diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,13989 @@
+{
+  "best_global_step": 488376,
+  "best_metric": 0.963716296062499,
+  "best_model_checkpoint": "/media/bramiozo/Storage2/DATA/NER/PII/hf_ai4privacy_multiclass/fold_0/checkpoint-488376",
+  "epoch": 6.0,
+  "eval_steps": 500,
+  "global_step": 488376,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0031451466762495352,
+      "grad_norm": 32.7712287902832,
+      "learning_rate": 1.99968671679198e-05,
+      "loss": 19.1473,
+      "step": 256
+    },
+    {
+      "epoch": 0.0062902933524990705,
+      "grad_norm": 42.52288055419922,
+      "learning_rate": 1.99937220502236e-05,
+      "loss": 7.8886,
+      "step": 512
+    },
+    {
+      "epoch": 0.009435440028748606,
+      "grad_norm": 336.1084899902344,
+      "learning_rate": 1.99905769325274e-05,
+      "loss": 4.2993,
+      "step": 768
+    },
+    {
+      "epoch": 0.012580586704998141,
+      "grad_norm": 115.21817779541016,
+      "learning_rate": 1.9987431814831197e-05,
+      "loss": 3.1289,
+      "step": 1024
+    },
+    {
+      "epoch": 0.015725733381247678,
+      "grad_norm": 42.798561096191406,
+      "learning_rate": 1.9984286697134996e-05,
+      "loss": 2.547,
+      "step": 1280
+    },
+    {
+      "epoch": 0.01887088005749721,
+      "grad_norm": 45.711795806884766,
+      "learning_rate": 1.9981141579438793e-05,
+      "loss": 2.1724,
+      "step": 1536
+    },
+    {
+      "epoch": 0.02201602673374675,
+      "grad_norm": 73.08441925048828,
+      "learning_rate": 1.9977996461742593e-05,
+      "loss": 2.2168,
+      "step": 1792
+    },
+    {
+      "epoch": 0.025161173409996282,
+      "grad_norm": 545.430908203125,
+      "learning_rate": 1.9974851344046392e-05,
+      "loss": 1.9585,
+      "step": 2048
+    },
+    {
+      "epoch": 0.02830632008624582,
+      "grad_norm": 63.36457824707031,
+      "learning_rate": 1.9971706226350192e-05,
+      "loss": 1.8719,
+      "step": 2304
+    },
+    {
+      "epoch": 0.031451466762495356,
+      "grad_norm": 537.1735229492188,
+      "learning_rate": 1.996856110865399e-05,
+      "loss": 1.6815,
+      "step": 2560
+    },
+    {
+      "epoch": 0.03459661343874489,
+      "grad_norm": 23.961442947387695,
+      "learning_rate": 1.996541599095779e-05,
+      "loss": 1.6802,
+      "step": 2816
+    },
+    {
+      "epoch": 0.03774176011499442,
+      "grad_norm": 51.69450759887695,
+      "learning_rate": 1.996227087326159e-05,
+      "loss": 1.5876,
+      "step": 3072
+    },
+    {
+      "epoch": 0.04088690679124396,
+      "grad_norm": 116.23927307128906,
+      "learning_rate": 1.9959125755565385e-05,
+      "loss": 1.5829,
+      "step": 3328
+    },
+    {
+      "epoch": 0.0440320534674935,
+      "grad_norm": 28.727214813232422,
+      "learning_rate": 1.9955980637869185e-05,
+      "loss": 1.4704,
+      "step": 3584
+    },
+    {
+      "epoch": 0.047177200143743034,
+      "grad_norm": 64.51776123046875,
+      "learning_rate": 1.9952835520172984e-05,
+      "loss": 1.3783,
+      "step": 3840
+    },
+    {
+      "epoch": 0.050322346819992564,
+      "grad_norm": 109.68195343017578,
+      "learning_rate": 1.994969040247678e-05,
+      "loss": 1.383,
+      "step": 4096
+    },
+    {
+      "epoch": 0.0534674934962421,
+      "grad_norm": 28.38018798828125,
+      "learning_rate": 1.994654528478058e-05,
+      "loss": 1.4144,
+      "step": 4352
+    },
+    {
+      "epoch": 0.05661264017249164,
+      "grad_norm": 42.122928619384766,
+      "learning_rate": 1.994340016708438e-05,
+      "loss": 1.3459,
+      "step": 4608
+    },
+    {
+      "epoch": 0.059757786848741175,
+      "grad_norm": 526.4948120117188,
+      "learning_rate": 1.9940255049388177e-05,
+      "loss": 1.341,
+      "step": 4864
+    },
+    {
+      "epoch": 0.06290293352499071,
+      "grad_norm": 23.917184829711914,
+      "learning_rate": 1.9937109931691977e-05,
+      "loss": 1.3926,
+      "step": 5120
+    },
+    {
+      "epoch": 0.06604808020124024,
+      "grad_norm": 15.578055381774902,
+      "learning_rate": 1.9933964813995776e-05,
+      "loss": 1.2834,
+      "step": 5376
+    },
+    {
+      "epoch": 0.06919322687748979,
+      "grad_norm": 10.9509916305542,
+      "learning_rate": 1.9930819696299576e-05,
+      "loss": 1.3267,
+      "step": 5632
+    },
+    {
+      "epoch": 0.07233837355373932,
+      "grad_norm": 48.65658950805664,
+      "learning_rate": 1.9927674578603373e-05,
+      "loss": 1.31,
+      "step": 5888
+    },
+    {
+      "epoch": 0.07548352022998885,
+      "grad_norm": 24.163028717041016,
+      "learning_rate": 1.9924529460907172e-05,
+      "loss": 1.2908,
+      "step": 6144
+    },
+    {
+      "epoch": 0.07862866690623839,
+      "grad_norm": 140.82144165039062,
+      "learning_rate": 1.992138434321097e-05,
+      "loss": 1.2996,
+      "step": 6400
+    },
+    {
+      "epoch": 0.08177381358248792,
+      "grad_norm": 25.045412063598633,
+      "learning_rate": 1.991823922551477e-05,
+      "loss": 1.3976,
+      "step": 6656
+    },
+    {
+      "epoch": 0.08491896025873746,
+      "grad_norm": 144.6173095703125,
+      "learning_rate": 1.9915094107818565e-05,
+      "loss": 1.2839,
+      "step": 6912
+    },
+    {
+      "epoch": 0.088064106934987,
+      "grad_norm": 13.057480812072754,
+      "learning_rate": 1.9911948990122368e-05,
+      "loss": 1.185,
+      "step": 7168
+    },
+    {
+      "epoch": 0.09120925361123652,
+      "grad_norm": 72.62432098388672,
+      "learning_rate": 1.9908803872426165e-05,
+      "loss": 1.2578,
+      "step": 7424
+    },
+    {
+      "epoch": 0.09435440028748607,
+      "grad_norm": 32.611907958984375,
+      "learning_rate": 1.9905658754729964e-05,
+      "loss": 1.2493,
+      "step": 7680
+    },
+    {
+      "epoch": 0.0974995469637356,
+      "grad_norm": 38.99099349975586,
+      "learning_rate": 1.990251363703376e-05,
+      "loss": 1.2033,
+      "step": 7936
+    },
+    {
+      "epoch": 0.10064469363998513,
+      "grad_norm": 159.5609893798828,
+      "learning_rate": 1.989936851933756e-05,
+      "loss": 1.2486,
+      "step": 8192
+    },
+    {
+      "epoch": 0.10378984031623467,
+      "grad_norm": 42.655113220214844,
+      "learning_rate": 1.989622340164136e-05,
+      "loss": 1.195,
+      "step": 8448
+    },
+    {
+      "epoch": 0.1069349869924842,
+      "grad_norm": 180.4906768798828,
+      "learning_rate": 1.989307828394516e-05,
+      "loss": 1.1549,
+      "step": 8704
+    },
+    {
+      "epoch": 0.11008013366873375,
+      "grad_norm": 183.71400451660156,
+      "learning_rate": 1.9889933166248957e-05,
+      "loss": 1.0973,
+      "step": 8960
+    },
+    {
+      "epoch": 0.11322528034498328,
+      "grad_norm": 22.372793197631836,
+      "learning_rate": 1.9886788048552757e-05,
+      "loss": 1.1579,
+      "step": 9216
+    },
+    {
+      "epoch": 0.1163704270212328,
+      "grad_norm": 23.469459533691406,
+      "learning_rate": 1.9883642930856556e-05,
+      "loss": 1.1536,
+      "step": 9472
+    },
+    {
+      "epoch": 0.11951557369748235,
+      "grad_norm": 130.577392578125,
+      "learning_rate": 1.9880497813160353e-05,
+      "loss": 1.1779,
+      "step": 9728
+    },
+    {
+      "epoch": 0.12266072037373188,
+      "grad_norm": 421.74359130859375,
+      "learning_rate": 1.9877352695464153e-05,
+      "loss": 1.1346,
+      "step": 9984
+    },
+    {
+      "epoch": 0.12580586704998142,
+      "grad_norm": 54.03622055053711,
+      "learning_rate": 1.987420757776795e-05,
+      "loss": 1.1281,
+      "step": 10240
+    },
+    {
+      "epoch": 0.12895101372623094,
+      "grad_norm": 41.32209777832031,
+      "learning_rate": 1.9871062460071752e-05,
+      "loss": 1.1412,
+      "step": 10496
+    },
+    {
+      "epoch": 0.13209616040248048,
+      "grad_norm": 18.568401336669922,
+      "learning_rate": 1.986791734237555e-05,
+      "loss": 1.1971,
+      "step": 10752
+    },
+    {
+      "epoch": 0.13524130707873003,
+      "grad_norm": 18.432279586791992,
+      "learning_rate": 1.986477222467935e-05,
+      "loss": 1.002,
+      "step": 11008
+    },
+    {
+      "epoch": 0.13838645375497957,
+      "grad_norm": 38.60474395751953,
+      "learning_rate": 1.9861627106983145e-05,
+      "loss": 1.0879,
+      "step": 11264
+    },
+    {
+      "epoch": 0.1415316004312291,
+      "grad_norm": 44.780517578125,
+      "learning_rate": 1.9858481989286945e-05,
+      "loss": 1.0693,
+      "step": 11520
+    },
+    {
+      "epoch": 0.14467674710747863,
+      "grad_norm": 2.2511823177337646,
+      "learning_rate": 1.985533687159074e-05,
+      "loss": 1.0275,
+      "step": 11776
+    },
+    {
+      "epoch": 0.14782189378372818,
+      "grad_norm": 72.40104675292969,
+      "learning_rate": 1.9852191753894544e-05,
+      "loss": 1.173,
+      "step": 12032
+    },
+    {
+      "epoch": 0.1509670404599777,
+      "grad_norm": 88.38461303710938,
+      "learning_rate": 1.984904663619834e-05,
+      "loss": 1.0833,
+      "step": 12288
+    },
+    {
+      "epoch": 0.15411218713622724,
+      "grad_norm": 91.39885711669922,
+      "learning_rate": 1.984590151850214e-05,
+      "loss": 1.1651,
+      "step": 12544
+    },
+    {
+      "epoch": 0.15725733381247678,
+      "grad_norm": 77.59071350097656,
+      "learning_rate": 1.9842756400805937e-05,
+      "loss": 1.04,
+      "step": 12800
+    },
+    {
+      "epoch": 0.1604024804887263,
+      "grad_norm": 72.61529541015625,
+      "learning_rate": 1.9839611283109737e-05,
+      "loss": 1.0219,
+      "step": 13056
+    },
+    {
+      "epoch": 0.16354762716497584,
+      "grad_norm": 62.259971618652344,
+      "learning_rate": 1.9836466165413536e-05,
+      "loss": 0.9948,
+      "step": 13312
+    },
+    {
+      "epoch": 0.16669277384122538,
+      "grad_norm": 110.95796966552734,
+      "learning_rate": 1.9833321047717333e-05,
+      "loss": 1.1103,
+      "step": 13568
+    },
+    {
+      "epoch": 0.16983792051747493,
+      "grad_norm": 27.00925064086914,
+      "learning_rate": 1.9830175930021133e-05,
+      "loss": 0.9939,
+      "step": 13824
+    },
+    {
+      "epoch": 0.17298306719372444,
+      "grad_norm": 7.5836076736450195,
+      "learning_rate": 1.9827030812324932e-05,
+      "loss": 0.9819,
+      "step": 14080
+    },
+    {
+      "epoch": 0.176128213869974,
+      "grad_norm": 314.71826171875,
+      "learning_rate": 1.9823885694628732e-05,
+      "loss": 1.0555,
+      "step": 14336
+    },
+    {
+      "epoch": 0.17927336054622353,
+      "grad_norm": 117.14920043945312,
+      "learning_rate": 1.982074057693253e-05,
+      "loss": 0.9894,
+      "step": 14592
+    },
+    {
+      "epoch": 0.18241850722247305,
+      "grad_norm": 115.92585754394531,
+      "learning_rate": 1.981759545923633e-05,
+      "loss": 1.0499,
+      "step": 14848
+    },
+    {
+      "epoch": 0.1855636538987226,
+      "grad_norm": 21.828420639038086,
+      "learning_rate": 1.9814450341540125e-05,
+      "loss": 1.0972,
+      "step": 15104
+    },
+    {
+      "epoch": 0.18870880057497214,
+      "grad_norm": 89.78067779541016,
+      "learning_rate": 1.9811305223843925e-05,
+      "loss": 1.0367,
+      "step": 15360
+    },
+    {
+      "epoch": 0.19185394725122165,
+      "grad_norm": 22.759586334228516,
+      "learning_rate": 1.9808160106147725e-05,
+      "loss": 1.0155,
+      "step": 15616
+    },
+    {
+      "epoch": 0.1949990939274712,
+      "grad_norm": 253.0929412841797,
+      "learning_rate": 1.9805014988451524e-05,
+      "loss": 0.9406,
+      "step": 15872
+    },
+    {
+      "epoch": 0.19814424060372074,
+      "grad_norm": 360.17901611328125,
+      "learning_rate": 1.980186987075532e-05,
+      "loss": 1.0057,
+      "step": 16128
+    },
+    {
+      "epoch": 0.20128938727997026,
+      "grad_norm": 2.00240421295166,
+      "learning_rate": 1.979872475305912e-05,
+      "loss": 1.0127,
+      "step": 16384
+    },
+    {
+      "epoch": 0.2044345339562198,
+      "grad_norm": 8.117267608642578,
+      "learning_rate": 1.9795579635362917e-05,
+      "loss": 1.0674,
+      "step": 16640
+    },
+    {
+      "epoch": 0.20757968063246934,
+      "grad_norm": 6.787088394165039,
+      "learning_rate": 1.9792434517666717e-05,
+      "loss": 0.9554,
+      "step": 16896
+    },
+    {
+      "epoch": 0.2107248273087189,
+      "grad_norm": 93.70729064941406,
+      "learning_rate": 1.9789289399970517e-05,
+      "loss": 1.0448,
+      "step": 17152
+    },
+    {
+      "epoch": 0.2138699739849684,
+      "grad_norm": 9.580097198486328,
+      "learning_rate": 1.9786144282274316e-05,
+      "loss": 0.9356,
+      "step": 17408
+    },
+    {
+      "epoch": 0.21701512066121795,
+      "grad_norm": 25.831886291503906,
+      "learning_rate": 1.9782999164578113e-05,
+      "loss": 0.9846,
+      "step": 17664
+    },
+    {
+      "epoch": 0.2201602673374675,
+      "grad_norm": 34.23259735107422,
+      "learning_rate": 1.9779854046881913e-05,
+      "loss": 0.9462,
+      "step": 17920
+    },
+    {
+      "epoch": 0.223305414013717,
+      "grad_norm": 41.13847351074219,
+      "learning_rate": 1.977670892918571e-05,
+      "loss": 0.9685,
+      "step": 18176
+    },
+    {
+      "epoch": 0.22645056068996655,
+      "grad_norm": 29.712556838989258,
+      "learning_rate": 1.977356381148951e-05,
+      "loss": 0.9716,
+      "step": 18432
+    },
+    {
+      "epoch": 0.2295957073662161,
+      "grad_norm": 120.3067398071289,
+      "learning_rate": 1.977041869379331e-05,
+      "loss": 1.0268,
+      "step": 18688
+    },
+    {
+      "epoch": 0.2327408540424656,
+      "grad_norm": 15.185142517089844,
+      "learning_rate": 1.976727357609711e-05,
+      "loss": 0.9563,
+      "step": 18944
+    },
+    {
+      "epoch": 0.23588600071871516,
+      "grad_norm": 48.46493911743164,
+      "learning_rate": 1.9764128458400905e-05,
+      "loss": 0.9826,
+      "step": 19200
+    },
+    {
+      "epoch": 0.2390311473949647,
+      "grad_norm": 21.320100784301758,
+      "learning_rate": 1.9760983340704705e-05,
+      "loss": 0.9483,
+      "step": 19456
+    },
+    {
+      "epoch": 0.24217629407121424,
+      "grad_norm": 39.47939682006836,
+      "learning_rate": 1.9757838223008504e-05,
+      "loss": 0.9643,
+      "step": 19712
+    },
+    {
+      "epoch": 0.24532144074746376,
+      "grad_norm": 57.34724807739258,
+      "learning_rate": 1.97546931053123e-05,
+      "loss": 1.0122,
+      "step": 19968
+    },
+    {
+      "epoch": 0.2484665874237133,
+      "grad_norm": 27.78389549255371,
+      "learning_rate": 1.97515479876161e-05,
+      "loss": 0.8975,
+      "step": 20224
+    },
+    {
+      "epoch": 0.25161173409996285,
+      "grad_norm": 33.786128997802734,
+      "learning_rate": 1.97484028699199e-05,
+      "loss": 0.9391,
+      "step": 20480
+    },
+    {
+      "epoch": 0.2547568807762124,
+      "grad_norm": 61.56987762451172,
+      "learning_rate": 1.97452577522237e-05,
+      "loss": 0.9268,
+      "step": 20736
+    },
+    {
+      "epoch": 0.2579020274524619,
+      "grad_norm": 37.85352325439453,
+      "learning_rate": 1.9742112634527497e-05,
+      "loss": 1.0033,
+      "step": 20992
+    },
+    {
+      "epoch": 0.2610471741287114,
+      "grad_norm": 17.906692504882812,
+      "learning_rate": 1.9738967516831296e-05,
+      "loss": 0.9741,
+      "step": 21248
+    },
+    {
+      "epoch": 0.26419232080496097,
+      "grad_norm": 21.666580200195312,
+      "learning_rate": 1.9735822399135093e-05,
+      "loss": 0.9962,
+      "step": 21504
+    },
+    {
+      "epoch": 0.2673374674812105,
+      "grad_norm": 13.210589408874512,
+      "learning_rate": 1.9732677281438893e-05,
+      "loss": 1.0377,
+      "step": 21760
+    },
+    {
+      "epoch": 0.27048261415746006,
+      "grad_norm": 25.927845001220703,
+      "learning_rate": 1.972953216374269e-05,
+      "loss": 0.8756,
+      "step": 22016
+    },
+    {
+      "epoch": 0.2736277608337096,
+      "grad_norm": 44.77389144897461,
+      "learning_rate": 1.9726387046046492e-05,
+      "loss": 1.0392,
+      "step": 22272
+    },
+    {
+      "epoch": 0.27677290750995914,
+      "grad_norm": 55.3034553527832,
+      "learning_rate": 1.972324192835029e-05,
+      "loss": 0.9849,
+      "step": 22528
+    },
+    {
+      "epoch": 0.27991805418620863,
+      "grad_norm": 43.52450942993164,
+      "learning_rate": 1.972009681065409e-05,
+      "loss": 0.9294,
+      "step": 22784
+    },
+    {
+      "epoch": 0.2830632008624582,
+      "grad_norm": 8.082113265991211,
+      "learning_rate": 1.9716951692957885e-05,
+      "loss": 0.9179,
+      "step": 23040
+    },
+    {
+      "epoch": 0.2862083475387077,
+      "grad_norm": 125.46047973632812,
+      "learning_rate": 1.9713806575261685e-05,
+      "loss": 0.9643,
+      "step": 23296
+    },
+    {
+      "epoch": 0.28935349421495726,
+      "grad_norm": 31.45306396484375,
+      "learning_rate": 1.9710661457565485e-05,
+      "loss": 1.0123,
+      "step": 23552
+    },
+    {
+      "epoch": 0.2924986408912068,
+      "grad_norm": 12.264525413513184,
+      "learning_rate": 1.9707516339869284e-05,
+      "loss": 0.9695,
+      "step": 23808
+    },
+    {
+      "epoch": 0.29564378756745635,
+      "grad_norm": 129.3623504638672,
+      "learning_rate": 1.970437122217308e-05,
+      "loss": 0.951,
+      "step": 24064
+    },
+    {
+      "epoch": 0.2987889342437059,
+      "grad_norm": 29.326669692993164,
+      "learning_rate": 1.970122610447688e-05,
+      "loss": 0.9386,
+      "step": 24320
+    },
+    {
+      "epoch": 0.3019340809199554,
+      "grad_norm": 110.87677764892578,
+      "learning_rate": 1.969808098678068e-05,
+      "loss": 0.9918,
+      "step": 24576
+    },
+    {
+      "epoch": 0.3050792275962049,
+      "grad_norm": 19.163772583007812,
+      "learning_rate": 1.9694935869084477e-05,
+      "loss": 0.8631,
+      "step": 24832
+    },
+    {
+      "epoch": 0.30822437427245447,
+      "grad_norm": 50.83871841430664,
+      "learning_rate": 1.9691790751388277e-05,
+      "loss": 0.9029,
+      "step": 25088
+    },
+    {
+      "epoch": 0.311369520948704,
+      "grad_norm": 183.310791015625,
+      "learning_rate": 1.9688645633692073e-05,
+      "loss": 0.9294,
+      "step": 25344
+    },
+    {
+      "epoch": 0.31451466762495356,
+      "grad_norm": 59.15810012817383,
+      "learning_rate": 1.9685500515995876e-05,
+      "loss": 0.9236,
+      "step": 25600
+    },
+    {
+      "epoch": 0.3176598143012031,
+      "grad_norm": 29.261056900024414,
+      "learning_rate": 1.9682355398299673e-05,
+      "loss": 0.9424,
+      "step": 25856
+    },
+    {
+      "epoch": 0.3208049609774526,
+      "grad_norm": 48.07884979248047,
+      "learning_rate": 1.9679210280603472e-05,
+      "loss": 0.8859,
+      "step": 26112
+    },
+    {
+      "epoch": 0.32395010765370214,
+      "grad_norm": 3.8921380043029785,
+      "learning_rate": 1.967606516290727e-05,
+      "loss": 0.9979,
+      "step": 26368
+    },
+    {
+      "epoch": 0.3270952543299517,
+      "grad_norm": 8.486544609069824,
+      "learning_rate": 1.967292004521107e-05,
+      "loss": 0.892,
+      "step": 26624
+    },
+    {
+      "epoch": 0.3302404010062012,
+      "grad_norm": 27.714935302734375,
+      "learning_rate": 1.9669774927514865e-05,
+      "loss": 0.963,
+      "step": 26880
+    },
+    {
+      "epoch": 0.33338554768245077,
+      "grad_norm": 22.318193435668945,
+      "learning_rate": 1.9666629809818668e-05,
+      "loss": 0.8974,
+      "step": 27136
+    },
+    {
+      "epoch": 0.3365306943587003,
+      "grad_norm": 60.13337326049805,
+      "learning_rate": 1.9663484692122465e-05,
+      "loss": 0.9975,
+      "step": 27392
+    },
+    {
+      "epoch": 0.33967584103494985,
+      "grad_norm": 1.0397473573684692,
+      "learning_rate": 1.9660339574426264e-05,
+      "loss": 0.8018,
+      "step": 27648
+    },
+    {
+      "epoch": 0.34282098771119934,
+      "grad_norm": 14.674808502197266,
+      "learning_rate": 1.965719445673006e-05,
+      "loss": 0.9126,
+      "step": 27904
+    },
+    {
+      "epoch": 0.3459661343874489,
+      "grad_norm": 40.55316162109375,
+      "learning_rate": 1.965404933903386e-05,
+      "loss": 1.0609,
+      "step": 28160
+    },
+    {
+      "epoch": 0.34911128106369843,
+      "grad_norm": 8.796433448791504,
+      "learning_rate": 1.965090422133766e-05,
+      "loss": 0.8669,
+      "step": 28416
+    },
+    {
+      "epoch": 0.352256427739948,
+      "grad_norm": 44.125709533691406,
+      "learning_rate": 1.9647759103641457e-05,
+      "loss": 1.0481,
+      "step": 28672
+    },
+    {
+      "epoch": 0.3554015744161975,
+      "grad_norm": 20.289310455322266,
+      "learning_rate": 1.9644613985945257e-05,
+      "loss": 0.8568,
+      "step": 28928
+    },
+    {
+      "epoch": 0.35854672109244706,
+      "grad_norm": 27.691425323486328,
+      "learning_rate": 1.9641468868249057e-05,
+      "loss": 0.898,
+      "step": 29184
+    },
+    {
+      "epoch": 0.36169186776869655,
+      "grad_norm": 38.798805236816406,
+      "learning_rate": 1.9638323750552853e-05,
+      "loss": 0.8529,
+      "step": 29440
+    },
+    {
+      "epoch": 0.3648370144449461,
+      "grad_norm": 11.752766609191895,
+      "learning_rate": 1.9635178632856653e-05,
+      "loss": 0.9455,
+      "step": 29696
+    },
+    {
+      "epoch": 0.36798216112119564,
+      "grad_norm": 49.48054885864258,
+      "learning_rate": 1.9632033515160453e-05,
+      "loss": 0.8921,
+      "step": 29952
+    },
+    {
+      "epoch": 0.3711273077974452,
+      "grad_norm": 35.214447021484375,
+      "learning_rate": 1.962888839746425e-05,
+      "loss": 0.8628,
+      "step": 30208
+    },
+    {
+      "epoch": 0.3742724544736947,
+      "grad_norm": 13.805008888244629,
+      "learning_rate": 1.962574327976805e-05,
+      "loss": 0.9266,
+      "step": 30464
+    },
+    {
+      "epoch": 0.37741760114994427,
+      "grad_norm": 39.77656555175781,
+      "learning_rate": 1.962259816207185e-05,
+      "loss": 0.922,
+      "step": 30720
+    },
+    {
+      "epoch": 0.3805627478261938,
+      "grad_norm": 22.713998794555664,
+      "learning_rate": 1.961945304437565e-05,
+      "loss": 0.9371,
+      "step": 30976
+    },
+    {
+      "epoch": 0.3837078945024433,
+      "grad_norm": 26.341171264648438,
+      "learning_rate": 1.9616307926679445e-05,
+      "loss": 0.8727,
+      "step": 31232
+    },
+    {
+      "epoch": 0.38685304117869285,
+      "grad_norm": 113.00861358642578,
+      "learning_rate": 1.9613162808983245e-05,
+      "loss": 0.9696,
+      "step": 31488
+    },
+    {
+      "epoch": 0.3899981878549424,
+      "grad_norm": 47.65739822387695,
+      "learning_rate": 1.961001769128704e-05,
+      "loss": 0.9086,
+      "step": 31744
+    },
+    {
+      "epoch": 0.39314333453119193,
+      "grad_norm": 18.171649932861328,
+      "learning_rate": 1.960687257359084e-05,
+      "loss": 0.9169,
+      "step": 32000
+    },
+    {
+      "epoch": 0.3962884812074415,
+      "grad_norm": 13.391619682312012,
+      "learning_rate": 1.960372745589464e-05,
+      "loss": 0.8035,
+      "step": 32256
+    },
+    {
+      "epoch": 0.399433627883691,
+      "grad_norm": 41.63272476196289,
+      "learning_rate": 1.960058233819844e-05,
+      "loss": 0.8812,
+      "step": 32512
+    },
+    {
+      "epoch": 0.4025787745599405,
+      "grad_norm": 66.81033325195312,
+      "learning_rate": 1.9597437220502237e-05,
+      "loss": 1.0155,
+      "step": 32768
+    },
+    {
+      "epoch": 0.40572392123619005,
+      "grad_norm": 226.61126708984375,
+      "learning_rate": 1.9594292102806037e-05,
+      "loss": 0.9564,
+      "step": 33024
+    },
+    {
+      "epoch": 0.4088690679124396,
+      "grad_norm": 68.86620330810547,
+      "learning_rate": 1.9591146985109833e-05,
+      "loss": 0.9151,
+      "step": 33280
+    },
+    {
+      "epoch": 0.41201421458868914,
+      "grad_norm": 13.57064151763916,
+      "learning_rate": 1.9588001867413633e-05,
+      "loss": 0.8516,
+      "step": 33536
+    },
+    {
+      "epoch": 0.4151593612649387,
+      "grad_norm": 67.34700775146484,
+      "learning_rate": 1.9584856749717433e-05,
+      "loss": 0.9247,
+      "step": 33792
+    },
+    {
+      "epoch": 0.41830450794118823,
+      "grad_norm": 13.842633247375488,
+      "learning_rate": 1.9581711632021232e-05,
+      "loss": 0.8434,
+      "step": 34048
+    },
+    {
+      "epoch": 0.4214496546174378,
+      "grad_norm": 86.75408935546875,
+      "learning_rate": 1.957856651432503e-05,
+      "loss": 0.9044,
+      "step": 34304
+    },
+    {
+      "epoch": 0.42459480129368726,
+      "grad_norm": 41.804054260253906,
+      "learning_rate": 1.957542139662883e-05,
+      "loss": 0.9508,
+      "step": 34560
+    },
+    {
+      "epoch": 0.4277399479699368,
+      "grad_norm": 193.1125030517578,
+      "learning_rate": 1.957227627893263e-05,
+      "loss": 0.8917,
+      "step": 34816
+    },
+    {
+      "epoch": 0.43088509464618635,
+      "grad_norm": 16.722246170043945,
+      "learning_rate": 1.9569131161236425e-05,
+      "loss": 0.8804,
+      "step": 35072
+    },
+    {
+      "epoch": 0.4340302413224359,
+      "grad_norm": 70.33397674560547,
+      "learning_rate": 1.9565986043540225e-05,
+      "loss": 0.8304,
+      "step": 35328
+    },
+    {
+      "epoch": 0.43717538799868544,
+      "grad_norm": 9.820713996887207,
+      "learning_rate": 1.9562840925844025e-05,
+      "loss": 0.9104,
+      "step": 35584
+    },
+    {
+      "epoch": 0.440320534674935,
+      "grad_norm": 32.16443634033203,
+      "learning_rate": 1.9559695808147824e-05,
+      "loss": 0.9114,
+      "step": 35840
+    },
+    {
+      "epoch": 0.44346568135118447,
+      "grad_norm": 7.383408069610596,
+      "learning_rate": 1.955655069045162e-05,
+      "loss": 0.9133,
+      "step": 36096
+    },
+    {
+      "epoch": 0.446610828027434,
+      "grad_norm": 4.049899578094482,
+      "learning_rate": 1.955340557275542e-05,
+      "loss": 0.8336,
+      "step": 36352
+    },
+    {
+      "epoch": 0.44975597470368356,
+      "grad_norm": 112.00721740722656,
+      "learning_rate": 1.9550260455059217e-05,
+      "loss": 0.9331,
+      "step": 36608
+    },
+    {
+      "epoch": 0.4529011213799331,
+      "grad_norm": 20.56984519958496,
+      "learning_rate": 1.9547115337363017e-05,
+      "loss": 0.7464,
+      "step": 36864
+    },
+    {
+      "epoch": 0.45604626805618265,
+      "grad_norm": 5.716044902801514,
+      "learning_rate": 1.9543970219666813e-05,
+      "loss": 0.9074,
+      "step": 37120
+    },
+    {
+      "epoch": 0.4591914147324322,
+      "grad_norm": 20.574831008911133,
+      "learning_rate": 1.9540825101970616e-05,
+      "loss": 0.8644,
+      "step": 37376
+    },
+    {
+      "epoch": 0.46233656140868173,
+      "grad_norm": 32.15157699584961,
+      "learning_rate": 1.9537679984274413e-05,
+      "loss": 0.8315,
+      "step": 37632
+    },
+    {
+      "epoch": 0.4654817080849312,
+      "grad_norm": 23.646764755249023,
+      "learning_rate": 1.9534534866578213e-05,
+      "loss": 0.8159,
+      "step": 37888
+    },
+    {
+      "epoch": 0.46862685476118077,
+      "grad_norm": 154.92152404785156,
+      "learning_rate": 1.953138974888201e-05,
+      "loss": 0.8697,
+      "step": 38144
+    },
+    {
+      "epoch": 0.4717720014374303,
+      "grad_norm": 20.131654739379883,
+      "learning_rate": 1.952824463118581e-05,
+      "loss": 0.829,
+      "step": 38400
+    },
+    {
+      "epoch": 0.47491714811367985,
+      "grad_norm": 23.77907371520996,
+      "learning_rate": 1.952509951348961e-05,
+      "loss": 0.8628,
+      "step": 38656
+    },
+    {
+      "epoch": 0.4780622947899294,
+      "grad_norm": 6.5311689376831055,
+      "learning_rate": 1.952195439579341e-05,
+      "loss": 0.8496,
+      "step": 38912
+    },
+    {
+      "epoch": 0.48120744146617894,
+      "grad_norm": 3.0753581523895264,
+      "learning_rate": 1.9518809278097205e-05,
+      "loss": 0.8172,
+      "step": 39168
+    },
+    {
+      "epoch": 0.4843525881424285,
+      "grad_norm": 28.618165969848633,
+      "learning_rate": 1.9515664160401005e-05,
+      "loss": 0.7458,
+      "step": 39424
+    },
+    {
+      "epoch": 0.487497734818678,
+      "grad_norm": 54.68094253540039,
+      "learning_rate": 1.9512519042704804e-05,
+      "loss": 0.8168,
+      "step": 39680
+    },
+    {
+      "epoch": 0.4906428814949275,
+      "grad_norm": 0.707294225692749,
+      "learning_rate": 1.95093739250086e-05,
+      "loss": 0.9658,
+      "step": 39936
+    },
+    {
+      "epoch": 0.49378802817117706,
+      "grad_norm": 20.761947631835938,
+      "learning_rate": 1.95062288073124e-05,
+      "loss": 0.7987,
+      "step": 40192
+    },
+    {
+      "epoch": 0.4969331748474266,
+      "grad_norm": 34.08995056152344,
+      "learning_rate": 1.9503083689616197e-05,
+      "loss": 0.7903,
+      "step": 40448
+    },
+    {
+      "epoch": 0.5000783215236762,
+      "grad_norm": 53.9713249206543,
+      "learning_rate": 1.9499938571919997e-05,
+      "loss": 0.8613,
+      "step": 40704
+    },
+    {
+      "epoch": 0.5032234681999257,
+      "grad_norm": 67.15633392333984,
+      "learning_rate": 1.9496793454223797e-05,
+      "loss": 0.8965,
+      "step": 40960
+    },
+    {
+      "epoch": 0.5063686148761752,
+      "grad_norm": 30.999074935913086,
+      "learning_rate": 1.9493648336527597e-05,
+      "loss": 0.9388,
+      "step": 41216
+    },
+    {
+      "epoch": 0.5095137615524248,
+      "grad_norm": 163.9065704345703,
+      "learning_rate": 1.9490503218831393e-05,
+      "loss": 0.9452,
+      "step": 41472
+    },
+    {
+      "epoch": 0.5126589082286743,
+      "grad_norm": 18.62287712097168,
+      "learning_rate": 1.9487358101135193e-05,
+      "loss": 0.8002,
+      "step": 41728
+    },
+    {
+      "epoch": 0.5158040549049238,
+      "grad_norm": 18.4913387298584,
+      "learning_rate": 1.948421298343899e-05,
+      "loss": 0.8824,
+      "step": 41984
+    },
+    {
+      "epoch": 0.5189492015811733,
+      "grad_norm": 43.897132873535156,
+      "learning_rate": 1.9481067865742792e-05,
+      "loss": 0.8757,
+      "step": 42240
+    },
+    {
+      "epoch": 0.5220943482574228,
+      "grad_norm": 66.5075454711914,
+      "learning_rate": 1.947792274804659e-05,
+      "loss": 0.8281,
+      "step": 42496
+    },
+    {
+      "epoch": 0.5252394949336724,
+      "grad_norm": 14.319323539733887,
+      "learning_rate": 1.947477763035039e-05,
+      "loss": 0.8343,
+      "step": 42752
+    },
+    {
+      "epoch": 0.5283846416099219,
+      "grad_norm": 79.51860046386719,
+      "learning_rate": 1.9471632512654185e-05,
+      "loss": 0.8381,
+      "step": 43008
+    },
+    {
+      "epoch": 0.5315297882861715,
+      "grad_norm": 9.574504852294922,
+      "learning_rate": 1.9468487394957985e-05,
+      "loss": 0.7763,
+      "step": 43264
+    },
+    {
+      "epoch": 0.534674934962421,
+      "grad_norm": 42.933433532714844,
+      "learning_rate": 1.946534227726178e-05,
+      "loss": 0.862,
+      "step": 43520
+    },
+    {
+      "epoch": 0.5378200816386706,
+      "grad_norm": 16.87189483642578,
+      "learning_rate": 1.946219715956558e-05,
+      "loss": 0.918,
+      "step": 43776
+    },
+    {
+      "epoch": 0.5409652283149201,
+      "grad_norm": 3.747722864151001,
+      "learning_rate": 1.945905204186938e-05,
+      "loss": 0.8631,
+      "step": 44032
+    },
+    {
+      "epoch": 0.5441103749911697,
+      "grad_norm": 93.89302825927734,
+      "learning_rate": 1.945590692417318e-05,
+      "loss": 0.9536,
+      "step": 44288
+    },
+    {
+      "epoch": 0.5472555216674192,
+      "grad_norm": 63.170291900634766,
+      "learning_rate": 1.9452761806476977e-05,
+      "loss": 0.7902,
+      "step": 44544
+    },
+    {
+      "epoch": 0.5504006683436687,
+      "grad_norm": 13.9670991897583,
+      "learning_rate": 1.9449616688780777e-05,
+      "loss": 0.9157,
+      "step": 44800
+    },
+    {
+      "epoch": 0.5535458150199183,
+      "grad_norm": 34.39453125,
+      "learning_rate": 1.9446471571084577e-05,
+      "loss": 0.7294,
+      "step": 45056
+    },
+    {
+      "epoch": 0.5566909616961677,
+      "grad_norm": 12.963068008422852,
+      "learning_rate": 1.9443326453388373e-05,
+      "loss": 0.8952,
+      "step": 45312
+    },
+    {
+      "epoch": 0.5598361083724173,
+      "grad_norm": 0.9055582284927368,
+      "learning_rate": 1.9440181335692173e-05,
+      "loss": 0.8127,
+      "step": 45568
+    },
+    {
+      "epoch": 0.5629812550486668,
+      "grad_norm": 22.191797256469727,
+      "learning_rate": 1.9437036217995973e-05,
+      "loss": 0.8524,
+      "step": 45824
+    },
+    {
+      "epoch": 0.5661264017249164,
+      "grad_norm": 27.787099838256836,
+      "learning_rate": 1.9433891100299772e-05,
+      "loss": 0.8681,
+      "step": 46080
+    },
+    {
+      "epoch": 0.5692715484011659,
+      "grad_norm": 53.42601013183594,
+      "learning_rate": 1.943074598260357e-05,
+      "loss": 0.8589,
+      "step": 46336
+    },
+    {
+      "epoch": 0.5724166950774154,
+      "grad_norm": 24.402891159057617,
+      "learning_rate": 1.942760086490737e-05,
+      "loss": 0.8512,
+      "step": 46592
+    },
+    {
+      "epoch": 0.575561841753665,
+      "grad_norm": 333.705322265625,
+      "learning_rate": 1.9424455747211165e-05,
+      "loss": 0.8897,
+      "step": 46848
+    },
+    {
+      "epoch": 0.5787069884299145,
+      "grad_norm": 0.12425386160612106,
+      "learning_rate": 1.9421310629514965e-05,
+      "loss": 0.8273,
+      "step": 47104
+    },
+    {
+      "epoch": 0.5818521351061641,
+      "grad_norm": 13.178757667541504,
+      "learning_rate": 1.9418165511818765e-05,
+      "loss": 0.8551,
+      "step": 47360
+    },
+    {
+      "epoch": 0.5849972817824136,
+      "grad_norm": 11.645500183105469,
+      "learning_rate": 1.9415020394122565e-05,
+      "loss": 0.8101,
+      "step": 47616
+    },
+    {
+      "epoch": 0.5881424284586632,
+      "grad_norm": 4.505655288696289,
+      "learning_rate": 1.941187527642636e-05,
+      "loss": 0.7911,
+      "step": 47872
+    },
+    {
+      "epoch": 0.5912875751349127,
+      "grad_norm": 32.914405822753906,
+      "learning_rate": 1.940873015873016e-05,
+      "loss": 0.8675,
+      "step": 48128
+    },
+    {
+      "epoch": 0.5944327218111622,
+      "grad_norm": 18.012737274169922,
+      "learning_rate": 1.9405585041033957e-05,
+      "loss": 0.8601,
+      "step": 48384
+    },
+    {
+      "epoch": 0.5975778684874118,
+      "grad_norm": 54.23798751831055,
+      "learning_rate": 1.9402439923337757e-05,
+      "loss": 0.7622,
+      "step": 48640
+    },
+    {
+      "epoch": 0.6007230151636612,
+      "grad_norm": 16.146764755249023,
+      "learning_rate": 1.9399294805641557e-05,
+      "loss": 0.8982,
+      "step": 48896
+    },
+    {
+      "epoch": 0.6038681618399108,
+      "grad_norm": 15.344330787658691,
+      "learning_rate": 1.9396149687945357e-05,
+      "loss": 0.8063,
+      "step": 49152
+    },
+    {
+      "epoch": 0.6070133085161603,
+      "grad_norm": 88.15870666503906,
+      "learning_rate": 1.9393004570249153e-05,
+      "loss": 0.8086,
+      "step": 49408
+    },
+    {
+      "epoch": 0.6101584551924099,
+      "grad_norm": 21.332048416137695,
+      "learning_rate": 1.9389859452552953e-05,
+      "loss": 0.8059,
+      "step": 49664
+    },
+    {
+      "epoch": 0.6133036018686594,
+      "grad_norm": 17.2802791595459,
+      "learning_rate": 1.9386714334856753e-05,
+      "loss": 0.7529,
+      "step": 49920
+    },
+    {
+      "epoch": 0.6164487485449089,
+      "grad_norm": 25.067615509033203,
+      "learning_rate": 1.938356921716055e-05,
+      "loss": 0.8333,
+      "step": 50176
+    },
+    {
+      "epoch": 0.6195938952211585,
+      "grad_norm": 3.615793228149414,
+      "learning_rate": 1.938042409946435e-05,
+      "loss": 0.7651,
+      "step": 50432
+    },
+    {
+      "epoch": 0.622739041897408,
+      "grad_norm": 53.881553649902344,
+      "learning_rate": 1.937727898176815e-05,
+      "loss": 0.7941,
+      "step": 50688
+    },
+    {
+      "epoch": 0.6258841885736576,
+      "grad_norm": 28.71366310119629,
+      "learning_rate": 1.937413386407195e-05,
+      "loss": 0.825,
+      "step": 50944
+    },
+    {
+      "epoch": 0.6290293352499071,
+      "grad_norm": 78.82333374023438,
+      "learning_rate": 1.9370988746375745e-05,
+      "loss": 0.8965,
+      "step": 51200
+    },
+    {
+      "epoch": 0.6321744819261567,
+      "grad_norm": 52.63657760620117,
+      "learning_rate": 1.9367843628679545e-05,
+      "loss": 0.8233,
+      "step": 51456
+    },
+    {
+      "epoch": 0.6353196286024062,
+      "grad_norm": 99.93594360351562,
+      "learning_rate": 1.936469851098334e-05,
+      "loss": 0.819,
+      "step": 51712
+    },
+    {
+      "epoch": 0.6384647752786557,
+      "grad_norm": 3.279900312423706,
+      "learning_rate": 1.936155339328714e-05,
+      "loss": 0.7395,
+      "step": 51968
+    },
+    {
+      "epoch": 0.6416099219549052,
+      "grad_norm": 19.55312728881836,
+      "learning_rate": 1.9358408275590937e-05,
+      "loss": 0.769,
+      "step": 52224
+    },
+    {
+      "epoch": 0.6447550686311547,
+      "grad_norm": 59.7762451171875,
+      "learning_rate": 1.935526315789474e-05,
+      "loss": 0.8845,
+      "step": 52480
+    },
+    {
+      "epoch": 0.6479002153074043,
+      "grad_norm": 102.45934295654297,
+      "learning_rate": 1.9352118040198537e-05,
+      "loss": 0.7528,
+      "step": 52736
+    },
+    {
+      "epoch": 0.6510453619836538,
+      "grad_norm": 3.297752857208252,
+      "learning_rate": 1.9348972922502337e-05,
+      "loss": 0.7772,
+      "step": 52992
+    },
+    {
+      "epoch": 0.6541905086599034,
+      "grad_norm": 134.29083251953125,
+      "learning_rate": 1.9345827804806133e-05,
+      "loss": 0.791,
+      "step": 53248
+    },
+    {
+      "epoch": 0.6573356553361529,
+      "grad_norm": 55.939613342285156,
+      "learning_rate": 1.9342682687109933e-05,
+      "loss": 0.7829,
+      "step": 53504
+    },
+    {
+      "epoch": 0.6604808020124024,
+      "grad_norm": 72.99360656738281,
+      "learning_rate": 1.9339537569413733e-05,
+      "loss": 0.7694,
+      "step": 53760
+    },
+    {
+      "epoch": 0.663625948688652,
+      "grad_norm": 4.768626689910889,
+      "learning_rate": 1.9336392451717533e-05,
+      "loss": 0.7946,
+      "step": 54016
+    },
+    {
+      "epoch": 0.6667710953649015,
+      "grad_norm": 193.68756103515625,
+      "learning_rate": 1.933324733402133e-05,
+      "loss": 0.8732,
+      "step": 54272
+    },
+    {
+      "epoch": 0.6699162420411511,
+      "grad_norm": 24.868032455444336,
+      "learning_rate": 1.933010221632513e-05,
+      "loss": 0.8656,
+      "step": 54528
+    },
+    {
+      "epoch": 0.6730613887174006,
+      "grad_norm": 26.19922637939453,
+      "learning_rate": 1.9326957098628925e-05,
+      "loss": 0.8286,
+      "step": 54784
+    },
+    {
+      "epoch": 0.6762065353936502,
+      "grad_norm": 48.13855743408203,
+      "learning_rate": 1.9323811980932725e-05,
+      "loss": 0.756,
+      "step": 55040
+    },
+    {
+      "epoch": 0.6793516820698997,
+      "grad_norm": 5.374435901641846,
+      "learning_rate": 1.9320666863236525e-05,
+      "loss": 0.8973,
+      "step": 55296
+    },
+    {
+      "epoch": 0.6824968287461491,
+      "grad_norm": 58.03184127807617,
+      "learning_rate": 1.931752174554032e-05,
+      "loss": 0.8593,
+      "step": 55552
+    },
+    {
+      "epoch": 0.6856419754223987,
+      "grad_norm": 71.15189361572266,
+      "learning_rate": 1.931437662784412e-05,
+      "loss": 0.7778,
+      "step": 55808
+    },
+    {
+      "epoch": 0.6887871220986482,
+      "grad_norm": 45.8195915222168,
+      "learning_rate": 1.931123151014792e-05,
+      "loss": 0.7534,
+      "step": 56064
+    },
+    {
+      "epoch": 0.6919322687748978,
+      "grad_norm": 46.83425521850586,
+      "learning_rate": 1.930808639245172e-05,
+      "loss": 0.885,
+      "step": 56320
+    },
+    {
+      "epoch": 0.6950774154511473,
+      "grad_norm": 13.252790451049805,
+      "learning_rate": 1.9304941274755517e-05,
+      "loss": 0.8796,
+      "step": 56576
+    },
+    {
+      "epoch": 0.6982225621273969,
+      "grad_norm": 22.359500885009766,
+      "learning_rate": 1.9301796157059317e-05,
+      "loss": 0.8106,
+      "step": 56832
+    },
+    {
+      "epoch": 0.7013677088036464,
+      "grad_norm": 0.6404379606246948,
+      "learning_rate": 1.9298651039363113e-05,
+      "loss": 0.8378,
+      "step": 57088
+    },
+    {
+      "epoch": 0.704512855479896,
+      "grad_norm": 81.59492492675781,
+      "learning_rate": 1.9295505921666916e-05,
+      "loss": 0.8661,
+      "step": 57344
+    },
+    {
+      "epoch": 0.7076580021561455,
+      "grad_norm": 32.51679611206055,
+      "learning_rate": 1.9292360803970713e-05,
+      "loss": 0.767,
+      "step": 57600
+    },
+    {
+      "epoch": 0.710803148832395,
+      "grad_norm": 63.804832458496094,
+      "learning_rate": 1.9289215686274513e-05,
+      "loss": 0.8529,
+      "step": 57856
+    },
+    {
+      "epoch": 0.7139482955086446,
+      "grad_norm": 7.646823883056641,
+      "learning_rate": 1.928607056857831e-05,
+      "loss": 0.7485,
+      "step": 58112
+    },
+    {
+      "epoch": 0.7170934421848941,
+      "grad_norm": 29.29694175720215,
+      "learning_rate": 1.928292545088211e-05,
+      "loss": 0.8579,
+      "step": 58368
+    },
+    {
+      "epoch": 0.7202385888611437,
+      "grad_norm": 71.68040466308594,
+      "learning_rate": 1.9279780333185905e-05,
+      "loss": 0.7033,
+      "step": 58624
+    },
+    {
+      "epoch": 0.7233837355373931,
+      "grad_norm": 41.10449981689453,
+      "learning_rate": 1.9276635215489705e-05,
+      "loss": 0.7697,
+      "step": 58880
+    },
+    {
+      "epoch": 0.7265288822136426,
+      "grad_norm": 23.591733932495117,
+      "learning_rate": 1.9273490097793505e-05,
+      "loss": 0.8284,
+      "step": 59136
+    },
+    {
+      "epoch": 0.7296740288898922,
+      "grad_norm": 53.10613250732422,
+      "learning_rate": 1.9270344980097305e-05,
+      "loss": 0.8086,
+      "step": 59392
+    },
+    {
+      "epoch": 0.7328191755661417,
+      "grad_norm": 176.76144409179688,
+      "learning_rate": 1.92671998624011e-05,
+      "loss": 0.8696,
+      "step": 59648
+    },
+    {
+      "epoch": 0.7359643222423913,
+      "grad_norm": 3.367558240890503,
+      "learning_rate": 1.92640547447049e-05,
+      "loss": 0.7546,
+      "step": 59904
+    },
+    {
+      "epoch": 0.7391094689186408,
+      "grad_norm": 34.550048828125,
+      "learning_rate": 1.92609096270087e-05,
+      "loss": 0.7839,
+      "step": 60160
+    },
+    {
+      "epoch": 0.7422546155948904,
+      "grad_norm": 26.86639976501465,
+      "learning_rate": 1.9257764509312497e-05,
+      "loss": 0.7813,
+      "step": 60416
+    },
+    {
+      "epoch": 0.7453997622711399,
+      "grad_norm": 16.04269027709961,
+      "learning_rate": 1.9254619391616297e-05,
+      "loss": 0.6915,
+      "step": 60672
+    },
+    {
+      "epoch": 0.7485449089473895,
+      "grad_norm": 117.5857162475586,
+      "learning_rate": 1.9251474273920097e-05,
+      "loss": 0.7728,
+      "step": 60928
+    },
+    {
+      "epoch": 0.751690055623639,
+      "grad_norm": 134.42347717285156,
+      "learning_rate": 1.9248329156223897e-05,
+      "loss": 0.813,
+      "step": 61184
+    },
+    {
+      "epoch": 0.7548352022998885,
+      "grad_norm": 23.08798599243164,
+      "learning_rate": 1.9245184038527693e-05,
+      "loss": 0.7486,
+      "step": 61440
+    },
+    {
+      "epoch": 0.7579803489761381,
+      "grad_norm": 49.708580017089844,
+      "learning_rate": 1.9242038920831493e-05,
+      "loss": 0.8235,
+      "step": 61696
+    },
+    {
+      "epoch": 0.7611254956523876,
+      "grad_norm": 4.749605178833008,
+      "learning_rate": 1.923889380313529e-05,
+      "loss": 0.6925,
+      "step": 61952
+    },
+    {
+      "epoch": 0.7642706423286371,
+      "grad_norm": 10.9595308303833,
+      "learning_rate": 1.923574868543909e-05,
+      "loss": 0.784,
+      "step": 62208
+    },
+    {
+      "epoch": 0.7674157890048866,
+      "grad_norm": 14.288825988769531,
+      "learning_rate": 1.923260356774289e-05,
+      "loss": 0.8013,
+      "step": 62464
+    },
+    {
+      "epoch": 0.7705609356811361,
+      "grad_norm": 5.931759834289551,
+      "learning_rate": 1.922945845004669e-05,
+      "loss": 0.6971,
+      "step": 62720
+    },
+    {
+      "epoch": 0.7737060823573857,
+      "grad_norm": 100.32432556152344,
+      "learning_rate": 1.9226313332350485e-05,
+      "loss": 0.818,
+      "step": 62976
+    },
+    {
+      "epoch": 0.7768512290336352,
+      "grad_norm": 19.227022171020508,
+      "learning_rate": 1.9223168214654285e-05,
+      "loss": 0.8415,
+      "step": 63232
+    },
+    {
+      "epoch": 0.7799963757098848,
+      "grad_norm": 26.44587516784668,
+      "learning_rate": 1.922002309695808e-05,
+      "loss": 0.7941,
+      "step": 63488
+    },
+    {
+      "epoch": 0.7831415223861343,
+      "grad_norm": 9.189475059509277,
+      "learning_rate": 1.921687797926188e-05,
+      "loss": 0.7718,
+      "step": 63744
+    },
+    {
+      "epoch": 0.7862866690623839,
+      "grad_norm": 73.33125305175781,
+      "learning_rate": 1.921373286156568e-05,
+      "loss": 0.7941,
+      "step": 64000
+    },
+    {
+      "epoch": 0.7894318157386334,
+      "grad_norm": 58.459716796875,
+      "learning_rate": 1.921058774386948e-05,
+      "loss": 0.821,
+      "step": 64256
+    },
+    {
+      "epoch": 0.792576962414883,
+      "grad_norm": 30.97920036315918,
+      "learning_rate": 1.9207442626173277e-05,
+      "loss": 0.809,
+      "step": 64512
+    },
+    {
+      "epoch": 0.7957221090911325,
+      "grad_norm": 23.164379119873047,
+      "learning_rate": 1.9204297508477077e-05,
+      "loss": 0.8261,
+      "step": 64768
+    },
+    {
+      "epoch": 0.798867255767382,
+      "grad_norm": 37.475868225097656,
+      "learning_rate": 1.9201152390780877e-05,
+      "loss": 0.7909,
+      "step": 65024
+    },
+    {
+      "epoch": 0.8020124024436316,
+      "grad_norm": 28.892026901245117,
+      "learning_rate": 1.9198007273084673e-05,
+      "loss": 0.7205,
+      "step": 65280
+    },
+    {
+      "epoch": 0.805157549119881,
+      "grad_norm": 16.609214782714844,
+      "learning_rate": 1.9194862155388473e-05,
+      "loss": 0.7656,
+      "step": 65536
+    },
+    {
+      "epoch": 0.8083026957961306,
+      "grad_norm": 187.5780487060547,
+      "learning_rate": 1.9191717037692273e-05,
+      "loss": 0.7644,
+      "step": 65792
+    },
+    {
+      "epoch": 0.8114478424723801,
+      "grad_norm": 18.108989715576172,
+      "learning_rate": 1.918857191999607e-05,
+      "loss": 0.8136,
+      "step": 66048
+    },
+    {
+      "epoch": 0.8145929891486297,
+      "grad_norm": 13.039497375488281,
+      "learning_rate": 1.918542680229987e-05,
+      "loss": 0.7235,
+      "step": 66304
+    },
+    {
+      "epoch": 0.8177381358248792,
+      "grad_norm": 45.3570442199707,
+      "learning_rate": 1.918228168460367e-05,
+      "loss": 0.7406,
+      "step": 66560
+    },
+    {
+      "epoch": 0.8208832825011287,
+      "grad_norm": 65.3688735961914,
+      "learning_rate": 1.9179136566907465e-05,
+      "loss": 0.7235,
+      "step": 66816
+    },
+    {
+      "epoch": 0.8240284291773783,
+      "grad_norm": 82.2174072265625,
+      "learning_rate": 1.9175991449211265e-05,
+      "loss": 0.8021,
+      "step": 67072
+    },
+    {
+      "epoch": 0.8271735758536278,
+      "grad_norm": 22.062021255493164,
+      "learning_rate": 1.9172846331515065e-05,
+      "loss": 0.7667,
+      "step": 67328
+    },
+    {
+      "epoch": 0.8303187225298774,
+      "grad_norm": 46.196449279785156,
+      "learning_rate": 1.9169701213818865e-05,
+      "loss": 0.6311,
+      "step": 67584
+    },
+    {
+      "epoch": 0.8334638692061269,
+      "grad_norm": 131.33926391601562,
+      "learning_rate": 1.916655609612266e-05,
+      "loss": 0.7878,
+      "step": 67840
+    },
+    {
+      "epoch": 0.8366090158823765,
+      "grad_norm": 26.568405151367188,
+      "learning_rate": 1.916341097842646e-05,
+      "loss": 0.7124,
+      "step": 68096
+    },
+    {
+      "epoch": 0.839754162558626,
+      "grad_norm": 38.21324157714844,
+      "learning_rate": 1.9160265860730257e-05,
+      "loss": 0.7942,
+      "step": 68352
+    },
+    {
+      "epoch": 0.8428993092348755,
+      "grad_norm": 110.54210662841797,
+      "learning_rate": 1.9157120743034057e-05,
+      "loss": 0.6726,
+      "step": 68608
+    },
+    {
+      "epoch": 0.846044455911125,
+      "grad_norm": 9.213847160339355,
+      "learning_rate": 1.9153975625337853e-05,
+      "loss": 0.7945,
+      "step": 68864
+    },
+    {
+      "epoch": 0.8491896025873745,
+      "grad_norm": 34.082672119140625,
+      "learning_rate": 1.9150830507641657e-05,
+      "loss": 0.7422,
+      "step": 69120
+    },
+    {
+      "epoch": 0.8523347492636241,
+      "grad_norm": 24.506345748901367,
+      "learning_rate": 1.9147685389945453e-05,
+      "loss": 0.6753,
+      "step": 69376
+    },
+    {
+      "epoch": 0.8554798959398736,
+      "grad_norm": 52.037025451660156,
+      "learning_rate": 1.9144540272249253e-05,
+      "loss": 0.8188,
+      "step": 69632
+    },
+    {
+      "epoch": 0.8586250426161232,
+      "grad_norm": 58.71159744262695,
+      "learning_rate": 1.914139515455305e-05,
+      "loss": 0.8136,
+      "step": 69888
+    },
+    {
+      "epoch": 0.8617701892923727,
+      "grad_norm": 0.6908836960792542,
+      "learning_rate": 1.913825003685685e-05,
+      "loss": 0.7699,
+      "step": 70144
+    },
+    {
+      "epoch": 0.8649153359686222,
+      "grad_norm": 3.233536958694458,
+      "learning_rate": 1.913510491916065e-05,
+      "loss": 0.7431,
+      "step": 70400
+    },
+    {
+      "epoch": 0.8680604826448718,
+      "grad_norm": 59.205318450927734,
+      "learning_rate": 1.913195980146445e-05,
+      "loss": 0.8062,
+      "step": 70656
+    },
+    {
+      "epoch": 0.8712056293211213,
+      "grad_norm": 23.355751037597656,
+      "learning_rate": 1.9128814683768245e-05,
+      "loss": 0.8281,
+      "step": 70912
+    },
+    {
+      "epoch": 0.8743507759973709,
+      "grad_norm": 21.780908584594727,
+      "learning_rate": 1.9125669566072045e-05,
+      "loss": 0.7592,
+      "step": 71168
+    },
+    {
+      "epoch": 0.8774959226736204,
+      "grad_norm": 47.90260696411133,
+      "learning_rate": 1.9122524448375845e-05,
+      "loss": 0.8206,
+      "step": 71424
+    },
+    {
+      "epoch": 0.88064106934987,
+      "grad_norm": 6.456783294677734,
+      "learning_rate": 1.911937933067964e-05,
+      "loss": 0.6849,
+      "step": 71680
+    },
+    {
+      "epoch": 0.8837862160261195,
+      "grad_norm": 11.97036361694336,
+      "learning_rate": 1.911623421298344e-05,
+      "loss": 0.7053,
+      "step": 71936
+    },
+    {
+      "epoch": 0.8869313627023689,
+      "grad_norm": 33.91883850097656,
+      "learning_rate": 1.9113089095287237e-05,
+      "loss": 0.7544,
+      "step": 72192
+    },
+    {
+      "epoch": 0.8900765093786185,
+      "grad_norm": 0.4971713125705719,
+      "learning_rate": 1.910994397759104e-05,
+      "loss": 0.7572,
+      "step": 72448
+    },
+    {
+      "epoch": 0.893221656054868,
+      "grad_norm": 47.002567291259766,
+      "learning_rate": 1.9106798859894837e-05,
+      "loss": 0.7579,
+      "step": 72704
+    },
+    {
+      "epoch": 0.8963668027311176,
+      "grad_norm": 63.87229537963867,
+      "learning_rate": 1.9103653742198637e-05,
+      "loss": 0.7395,
+      "step": 72960
+    },
+    {
+      "epoch": 0.8995119494073671,
+      "grad_norm": 47.33967590332031,
+      "learning_rate": 1.9100508624502433e-05,
+      "loss": 0.7718,
+      "step": 73216
+    },
+    {
+      "epoch": 0.9026570960836167,
+      "grad_norm": 7.953878879547119,
+      "learning_rate": 1.9097363506806233e-05,
+      "loss": 0.7522,
+      "step": 73472
+    },
+    {
+      "epoch": 0.9058022427598662,
+      "grad_norm": 64.82489013671875,
+      "learning_rate": 1.909421838911003e-05,
+      "loss": 0.7384,
+      "step": 73728
+    },
+    {
+      "epoch": 0.9089473894361157,
+      "grad_norm": 35.73405075073242,
+      "learning_rate": 1.909107327141383e-05,
+      "loss": 0.8237,
+      "step": 73984
+    },
+    {
+      "epoch": 0.9120925361123653,
+      "grad_norm": 148.7559814453125,
+      "learning_rate": 1.908792815371763e-05,
+      "loss": 0.7687,
+      "step": 74240
+    },
+    {
+      "epoch": 0.9152376827886148,
+      "grad_norm": 7.517806053161621,
+      "learning_rate": 1.908478303602143e-05,
+      "loss": 0.7234,
+      "step": 74496
+    },
+    {
+      "epoch": 0.9183828294648644,
+      "grad_norm": 2.0128366947174072,
+      "learning_rate": 1.9081637918325225e-05,
+      "loss": 0.7649,
+      "step": 74752
+    },
+    {
+      "epoch": 0.9215279761411139,
+      "grad_norm": 37.99541091918945,
+      "learning_rate": 1.9078492800629025e-05,
+      "loss": 0.7592,
+      "step": 75008
+    },
+    {
+      "epoch": 0.9246731228173635,
+      "grad_norm": 1.396522045135498,
+      "learning_rate": 1.9075347682932825e-05,
+      "loss": 0.7497,
+      "step": 75264
+    },
+    {
+      "epoch": 0.9278182694936129,
+      "grad_norm": 157.14682006835938,
+      "learning_rate": 1.907220256523662e-05,
+      "loss": 0.7398,
+      "step": 75520
+    },
+    {
+      "epoch": 0.9309634161698624,
+      "grad_norm": 2.272528886795044,
+      "learning_rate": 1.906905744754042e-05,
+      "loss": 0.7649,
+      "step": 75776
+    },
+    {
+      "epoch": 0.934108562846112,
+      "grad_norm": 15.956674575805664,
+      "learning_rate": 1.906591232984422e-05,
+      "loss": 0.8545,
+      "step": 76032
+    },
+    {
+      "epoch": 0.9372537095223615,
+      "grad_norm": 51.1855354309082,
+      "learning_rate": 1.906276721214802e-05,
+      "loss": 0.8055,
+      "step": 76288
+    },
+    {
+      "epoch": 0.9403988561986111,
+      "grad_norm": 81.28685760498047,
+      "learning_rate": 1.9059622094451817e-05,
+      "loss": 0.6968,
+      "step": 76544
+    },
+    {
+      "epoch": 0.9435440028748606,
+      "grad_norm": 11.084750175476074,
+      "learning_rate": 1.9056476976755617e-05,
+      "loss": 0.7754,
+      "step": 76800
+    },
+    {
+      "epoch": 0.9466891495511102,
+      "grad_norm": 26.921661376953125,
+      "learning_rate": 1.9053331859059413e-05,
+      "loss": 0.7317,
+      "step": 77056
+    },
+    {
+      "epoch": 0.9498342962273597,
+      "grad_norm": 102.69436645507812,
+      "learning_rate": 1.9050186741363213e-05,
+      "loss": 0.7334,
+      "step": 77312
+    },
+    {
+      "epoch": 0.9529794429036093,
+      "grad_norm": 132.38275146484375,
+      "learning_rate": 1.9047041623667013e-05,
+      "loss": 0.7844,
+      "step": 77568
+    },
+    {
+      "epoch": 0.9561245895798588,
+      "grad_norm": 34.51591491699219,
+      "learning_rate": 1.9043896505970813e-05,
+      "loss": 0.7996,
+      "step": 77824
+    },
+    {
+      "epoch": 0.9592697362561083,
+      "grad_norm": 27.336862564086914,
+      "learning_rate": 1.904075138827461e-05,
+      "loss": 0.7902,
+      "step": 78080
+    },
+    {
+      "epoch": 0.9624148829323579,
+      "grad_norm": 1225.2789306640625,
+      "learning_rate": 1.903760627057841e-05,
+      "loss": 0.7081,
+      "step": 78336
+    },
+    {
+      "epoch": 0.9655600296086074,
+      "grad_norm": 9.193857192993164,
+      "learning_rate": 1.9034461152882205e-05,
+      "loss": 0.7795,
+      "step": 78592
+    },
+    {
+      "epoch": 0.968705176284857,
+      "grad_norm": 39.14468002319336,
+      "learning_rate": 1.9031316035186005e-05,
+      "loss": 0.7788,
+      "step": 78848
+    },
+    {
+      "epoch": 0.9718503229611064,
+      "grad_norm": 7.461280345916748,
+      "learning_rate": 1.9028170917489805e-05,
+      "loss": 0.7368,
+      "step": 79104
+    },
+    {
+      "epoch": 0.974995469637356,
+      "grad_norm": 114.1192855834961,
+      "learning_rate": 1.9025025799793605e-05,
+      "loss": 0.7676,
+      "step": 79360
+    },
+    {
+      "epoch": 0.9781406163136055,
+      "grad_norm": 65.84223175048828,
+      "learning_rate": 1.90218806820974e-05,
+      "loss": 0.8094,
+      "step": 79616
+    },
+    {
+      "epoch": 0.981285762989855,
+      "grad_norm": 38.33774185180664,
+      "learning_rate": 1.90187355644012e-05,
+      "loss": 0.837,
+      "step": 79872
+    },
+    {
+      "epoch": 0.9844309096661046,
+      "grad_norm": 33.37348175048828,
+      "learning_rate": 1.9015590446704997e-05,
+      "loss": 0.7374,
+      "step": 80128
+    },
+    {
+      "epoch": 0.9875760563423541,
+      "grad_norm": 2.5657684803009033,
+      "learning_rate": 1.9012445329008797e-05,
+      "loss": 0.8291,
+      "step": 80384
+    },
+    {
+      "epoch": 0.9907212030186037,
+      "grad_norm": 102.24066162109375,
+      "learning_rate": 1.9009300211312597e-05,
+      "loss": 0.7558,
+      "step": 80640
+    },
+    {
+      "epoch": 0.9938663496948532,
+      "grad_norm": 216.89869689941406,
+      "learning_rate": 1.9006155093616397e-05,
+      "loss": 0.7639,
+      "step": 80896
+    },
+    {
+      "epoch": 0.9970114963711028,
+      "grad_norm": 38.7929573059082,
+      "learning_rate": 1.9003009975920193e-05,
+      "loss": 0.7199,
+      "step": 81152
+    },
+    {
+      "epoch": 1.0,
+      "eval_AGE": {
+        "f1": 0.9437821446621281,
+        "number": 32192,
+        "precision": 0.9049085000855139,
+        "recall": 0.9861456262425448
+      },
+      "eval_BUILDINGNUM": {
+        "f1": 0.963967009194159,
+        "number": 28889,
+        "precision": 0.9418408798176955,
+        "recall": 0.9871577417009935
+      },
+      "eval_CITY": {
+        "f1": 0.955791177275594,
+        "number": 41669,
+        "precision": 0.947951090548579,
+        "recall": 0.9637620293263577
+      },
+      "eval_DATE": {
+        "f1": 0.9961829472087417,
+        "number": 80940,
+        "precision": 0.9944230366749972,
+        "recall": 0.9979490980973561
+      },
+      "eval_EMAIL": {
+        "f1": 0.991642915801169,
+        "number": 51445,
+        "precision": 0.9857856319631195,
+        "recall": 0.9975702206239674
+      },
+      "eval_GENDER": {
+        "f1": 0.9674583769938293,
+        "number": 17457,
+        "precision": 0.9834309722468785,
+        "recall": 0.9519963338488858
+      },
+      "eval_GIVENNAME": {
+        "f1": 0.8971823561001325,
+        "number": 100350,
+        "precision": 0.88725504526364,
+        "recall": 0.9073343298455406
+      },
+      "eval_IDCARDNUM": {
+        "f1": 0.8694996228312799,
+        "number": 22068,
+        "precision": 0.8086997193638915,
+        "recall": 0.9401848830886351
+      },
+      "eval_PASSPORTNUM": {
+        "f1": 0.8923008290271014,
+        "number": 12959,
+        "precision": 0.8609038737446199,
+        "recall": 0.9260745427887954
+      },
+      "eval_SEX": {
+        "f1": 0.9593965975544924,
+        "number": 14570,
+        "precision": 0.9298595903645498,
+        "recall": 0.9908716540837337
+      },
+      "eval_SOCIALNUM": {
+        "f1": 0.8789161211752169,
+        "number": 15327,
+        "precision": 0.823952505993835,
+        "recall": 0.9417368043322242
+      },
+      "eval_STREET": {
+        "f1": 0.9657137687714855,
+        "number": 32803,
+        "precision": 0.955371259807285,
+        "recall": 0.9762826570740482
+      },
+      "eval_SURNAME": {
+        "f1": 0.8859171361736307,
+        "number": 85192,
+        "precision": 0.8849573070030571,
+        "recall": 0.8868790496760259
+      },
+      "eval_TELEPHONENUM": {
+        "f1": 0.987563694053429,
+        "number": 36941,
+        "precision": 0.9810086008867995,
+        "recall": 0.9942069786957581
+      },
+      "eval_ZIPCODE": {
+        "f1": 0.9619003298059607,
+        "number": 24346,
+        "precision": 0.9478805279738406,
+        "recall": 0.9763410827240614
+      },
+      "eval_overall_accuracy": 0.9890704747822591,
+      "eval_overall_f1": 0.9406098538209325,
+      "eval_overall_precision": 0.9272663667352766,
+      "eval_overall_recall": 0.9543429769504378,
+      "eval_runtime": 3737.1616,
+      "eval_samples_per_second": 87.12,
+      "eval_steps_per_second": 43.56,
+      "step": 81396
+    },
+    {
+      "epoch": 1.0001474287504493,
+      "grad_norm": 7.581768035888672,
+      "learning_rate": 1.8999864858223993e-05,
+      "loss": 0.7444,
+      "step": 81408
+    },
+    {
+      "epoch": 1.0032925754266988,
+      "grad_norm": 7.458587646484375,
+      "learning_rate": 1.8996719740527793e-05,
+      "loss": 0.6508,
+      "step": 81664
+    },
+    {
+      "epoch": 1.0064377221029484,
+      "grad_norm": 3.7170307636260986,
+      "learning_rate": 1.899357462283159e-05,
+      "loss": 0.6622,
+      "step": 81920
+    },
+    {
+      "epoch": 1.009582868779198,
+      "grad_norm": 2.339949607849121,
+      "learning_rate": 1.899042950513539e-05,
+      "loss": 0.6732,
+      "step": 82176
+    },
+    {
+      "epoch": 1.0127280154554474,
+      "grad_norm": 3.0435664653778076,
+      "learning_rate": 1.898728438743919e-05,
+      "loss": 0.6653,
+      "step": 82432
+    },
+    {
+      "epoch": 1.015873162131697,
+      "grad_norm": 258.94476318359375,
+      "learning_rate": 1.898413926974299e-05,
+      "loss": 0.6856,
+      "step": 82688
+    },
+    {
+      "epoch": 1.0190183088079463,
+      "grad_norm": 230.9947052001953,
+      "learning_rate": 1.8980994152046785e-05,
+      "loss": 0.6303,
+      "step": 82944
+    },
+    {
+      "epoch": 1.0221634554841958,
+      "grad_norm": 37.18260192871094,
+      "learning_rate": 1.8977849034350585e-05,
+      "loss": 0.6865,
+      "step": 83200
+    },
+    {
+      "epoch": 1.0253086021604454,
+      "grad_norm": 72.14546966552734,
+      "learning_rate": 1.897470391665438e-05,
+      "loss": 0.6349,
+      "step": 83456
+    },
+    {
+      "epoch": 1.028453748836695,
+      "grad_norm": 30.97728157043457,
+      "learning_rate": 1.897155879895818e-05,
+      "loss": 0.7054,
+      "step": 83712
+    },
+    {
+      "epoch": 1.0315988955129445,
+      "grad_norm": 20.211627960205078,
+      "learning_rate": 1.8968413681261977e-05,
+      "loss": 0.611,
+      "step": 83968
+    },
+    {
+      "epoch": 1.034744042189194,
+      "grad_norm": 40.315162658691406,
+      "learning_rate": 1.896526856356578e-05,
+      "loss": 0.6506,
+      "step": 84224
+    },
+    {
+      "epoch": 1.0378891888654436,
+      "grad_norm": 30.39170265197754,
+      "learning_rate": 1.8962123445869577e-05,
+      "loss": 0.6228,
+      "step": 84480
+    },
+    {
+      "epoch": 1.0410343355416931,
+      "grad_norm": 39.049537658691406,
+      "learning_rate": 1.8958978328173377e-05,
+      "loss": 0.7732,
+      "step": 84736
+    },
+    {
+      "epoch": 1.0441794822179427,
+      "grad_norm": 5.290384769439697,
+      "learning_rate": 1.8955833210477173e-05,
+      "loss": 0.7261,
+      "step": 84992
+    },
+    {
+      "epoch": 1.0473246288941922,
+      "grad_norm": 9.724326133728027,
+      "learning_rate": 1.8952688092780973e-05,
+      "loss": 0.6605,
+      "step": 85248
+    },
+    {
+      "epoch": 1.0504697755704417,
+      "grad_norm": 4.473785877227783,
+      "learning_rate": 1.8949542975084773e-05,
+      "loss": 0.6962,
+      "step": 85504
+    },
+    {
+      "epoch": 1.0536149222466913,
+      "grad_norm": 33.375221252441406,
+      "learning_rate": 1.8946397857388573e-05,
+      "loss": 0.6472,
+      "step": 85760
+    },
+    {
+      "epoch": 1.0567600689229408,
+      "grad_norm": 16.993270874023438,
+      "learning_rate": 1.894325273969237e-05,
+      "loss": 0.6813,
+      "step": 86016
+    },
+    {
+      "epoch": 1.0599052155991904,
+      "grad_norm": 53.01995086669922,
+      "learning_rate": 1.894010762199617e-05,
+      "loss": 0.7192,
+      "step": 86272
+    },
+    {
+      "epoch": 1.06305036227544,
+      "grad_norm": 51.68879699707031,
+      "learning_rate": 1.893696250429997e-05,
+      "loss": 0.7051,
+      "step": 86528
+    },
+    {
+      "epoch": 1.0661955089516895,
+      "grad_norm": 93.0103530883789,
+      "learning_rate": 1.8933817386603765e-05,
+      "loss": 0.6045,
+      "step": 86784
+    },
+    {
+      "epoch": 1.069340655627939,
+      "grad_norm": 5.482593059539795,
+      "learning_rate": 1.8930672268907565e-05,
+      "loss": 0.6694,
+      "step": 87040
+    },
+    {
+      "epoch": 1.0724858023041886,
+      "grad_norm": 26.17112922668457,
+      "learning_rate": 1.892752715121136e-05,
+      "loss": 0.6775,
+      "step": 87296
+    },
+    {
+      "epoch": 1.075630948980438,
+      "grad_norm": 89.9710922241211,
+      "learning_rate": 1.8924382033515165e-05,
+      "loss": 0.6526,
+      "step": 87552
+    },
+    {
+      "epoch": 1.0787760956566876,
+      "grad_norm": 12.555326461791992,
+      "learning_rate": 1.892123691581896e-05,
+      "loss": 0.6894,
+      "step": 87808
+    },
+    {
+      "epoch": 1.0819212423329372,
+      "grad_norm": 35.603553771972656,
+      "learning_rate": 1.891809179812276e-05,
+      "loss": 0.6822,
+      "step": 88064
+    },
+    {
+      "epoch": 1.0850663890091867,
+      "grad_norm": 51.86898422241211,
+      "learning_rate": 1.8914946680426557e-05,
+      "loss": 0.707,
+      "step": 88320
+    },
+    {
+      "epoch": 1.0882115356854363,
+      "grad_norm": 21.16473388671875,
+      "learning_rate": 1.8911801562730357e-05,
+      "loss": 0.6882,
+      "step": 88576
+    },
+    {
+      "epoch": 1.0913566823616858,
+      "grad_norm": 0.03948088735342026,
+      "learning_rate": 1.8908656445034153e-05,
+      "loss": 0.6757,
+      "step": 88832
+    },
+    {
+      "epoch": 1.0945018290379354,
+      "grad_norm": 2.8939571380615234,
+      "learning_rate": 1.8905511327337957e-05,
+      "loss": 0.609,
+      "step": 89088
+    },
+    {
+      "epoch": 1.0976469757141847,
+      "grad_norm": 154.57899475097656,
+      "learning_rate": 1.8902366209641753e-05,
+      "loss": 0.636,
+      "step": 89344
+    },
+    {
+      "epoch": 1.1007921223904344,
+      "grad_norm": 28.568700790405273,
+      "learning_rate": 1.8899221091945553e-05,
+      "loss": 0.6269,
+      "step": 89600
+    },
+    {
+      "epoch": 1.1039372690666838,
+      "grad_norm": 10.655451774597168,
+      "learning_rate": 1.889607597424935e-05,
+      "loss": 0.6368,
+      "step": 89856
+    },
+    {
+      "epoch": 1.1070824157429333,
+      "grad_norm": 1.2805752754211426,
+      "learning_rate": 1.889293085655315e-05,
+      "loss": 0.6713,
+      "step": 90112
+    },
+    {
+      "epoch": 1.1102275624191829,
+      "grad_norm": 1.600464940071106,
+      "learning_rate": 1.888978573885695e-05,
+      "loss": 0.6657,
+      "step": 90368
+    },
+    {
+      "epoch": 1.1133727090954324,
+      "grad_norm": 10.881330490112305,
+      "learning_rate": 1.8886640621160745e-05,
+      "loss": 0.7479,
+      "step": 90624
+    },
+    {
+      "epoch": 1.116517855771682,
+      "grad_norm": 35.77056884765625,
+      "learning_rate": 1.8883495503464545e-05,
+      "loss": 0.6778,
+      "step": 90880
+    },
+    {
+      "epoch": 1.1196630024479315,
+      "grad_norm": 70.57062530517578,
+      "learning_rate": 1.8880350385768345e-05,
+      "loss": 0.7036,
+      "step": 91136
+    },
+    {
+      "epoch": 1.122808149124181,
+      "grad_norm": 8.889589309692383,
+      "learning_rate": 1.887720526807214e-05,
+      "loss": 0.7747,
+      "step": 91392
+    },
+    {
+      "epoch": 1.1259532958004306,
+      "grad_norm": 33.128475189208984,
+      "learning_rate": 1.887406015037594e-05,
+      "loss": 0.6402,
+      "step": 91648
+    },
+    {
+      "epoch": 1.1290984424766801,
+      "grad_norm": 2.7361397743225098,
+      "learning_rate": 1.887091503267974e-05,
+      "loss": 0.7461,
+      "step": 91904
+    },
+    {
+      "epoch": 1.1322435891529297,
+      "grad_norm": 1.889805793762207,
+      "learning_rate": 1.8867769914983537e-05,
+      "loss": 0.696,
+      "step": 92160
+    },
+    {
+      "epoch": 1.1353887358291792,
+      "grad_norm": 53.63256072998047,
+      "learning_rate": 1.8864624797287337e-05,
+      "loss": 0.7495,
+      "step": 92416
+    },
+    {
+      "epoch": 1.1385338825054288,
+      "grad_norm": 37.536956787109375,
+      "learning_rate": 1.8861479679591137e-05,
+      "loss": 0.6901,
+      "step": 92672
+    },
+    {
+      "epoch": 1.1416790291816783,
+      "grad_norm": 28.29387664794922,
+      "learning_rate": 1.8858334561894937e-05,
+      "loss": 0.7264,
+      "step": 92928
+    },
+    {
+      "epoch": 1.1448241758579278,
+      "grad_norm": 309.3240966796875,
+      "learning_rate": 1.8855189444198733e-05,
+      "loss": 0.6539,
+      "step": 93184
+    },
+    {
+      "epoch": 1.1479693225341774,
+      "grad_norm": 43.505828857421875,
+      "learning_rate": 1.8852044326502533e-05,
+      "loss": 0.6069,
+      "step": 93440
+    },
+    {
+      "epoch": 1.151114469210427,
+      "grad_norm": 9.774971961975098,
+      "learning_rate": 1.884889920880633e-05,
+      "loss": 0.6669,
+      "step": 93696
+    },
+    {
+      "epoch": 1.1542596158866765,
+      "grad_norm": 45.66516876220703,
+      "learning_rate": 1.884575409111013e-05,
+      "loss": 0.7843,
+      "step": 93952
+    },
+    {
+      "epoch": 1.157404762562926,
+      "grad_norm": 15.896878242492676,
+      "learning_rate": 1.884260897341393e-05,
+      "loss": 0.7548,
+      "step": 94208
+    },
+    {
+      "epoch": 1.1605499092391756,
+      "grad_norm": 11.016946792602539,
+      "learning_rate": 1.883946385571773e-05,
+      "loss": 0.7924,
+      "step": 94464
+    },
+    {
+      "epoch": 1.163695055915425,
+      "grad_norm": 13.065671920776367,
+      "learning_rate": 1.8836318738021525e-05,
+      "loss": 0.7096,
+      "step": 94720
+    },
+    {
+      "epoch": 1.1668402025916746,
+      "grad_norm": 15.612651824951172,
+      "learning_rate": 1.8833173620325325e-05,
+      "loss": 0.6359,
+      "step": 94976
+    },
+    {
+      "epoch": 1.1699853492679242,
+      "grad_norm": 7.437082767486572,
+      "learning_rate": 1.883002850262912e-05,
+      "loss": 0.6947,
+      "step": 95232
+    },
+    {
+      "epoch": 1.1731304959441737,
+      "grad_norm": 90.36182403564453,
+      "learning_rate": 1.882688338493292e-05,
+      "loss": 0.6921,
+      "step": 95488
+    },
+    {
+      "epoch": 1.1762756426204233,
+      "grad_norm": 33.56642532348633,
+      "learning_rate": 1.882373826723672e-05,
+      "loss": 0.7084,
+      "step": 95744
+    },
+    {
+      "epoch": 1.1794207892966728,
+      "grad_norm": 56.74701690673828,
+      "learning_rate": 1.882059314954052e-05,
+      "loss": 0.6604,
+      "step": 96000
+    },
+    {
+      "epoch": 1.1825659359729221,
+      "grad_norm": 250.30209350585938,
+      "learning_rate": 1.8817448031844317e-05,
+      "loss": 0.6781,
+      "step": 96256
+    },
+    {
+      "epoch": 1.185711082649172,
+      "grad_norm": 114.74589538574219,
+      "learning_rate": 1.8814302914148117e-05,
+      "loss": 0.7663,
+      "step": 96512
+    },
+    {
+      "epoch": 1.1888562293254212,
+      "grad_norm": 1.8678464889526367,
+      "learning_rate": 1.8811157796451917e-05,
+      "loss": 0.7333,
+      "step": 96768
+    },
+    {
+      "epoch": 1.1920013760016708,
+      "grad_norm": 8.966241836547852,
+      "learning_rate": 1.8808012678755713e-05,
+      "loss": 0.6599,
+      "step": 97024
+    },
+    {
+      "epoch": 1.1951465226779203,
+      "grad_norm": 2.2832465171813965,
+      "learning_rate": 1.8804867561059513e-05,
+      "loss": 0.6934,
+      "step": 97280
+    },
+    {
+      "epoch": 1.1982916693541699,
+      "grad_norm": 32.91651916503906,
+      "learning_rate": 1.8801722443363313e-05,
+      "loss": 0.7224,
+      "step": 97536
+    },
+    {
+      "epoch": 1.2014368160304194,
+      "grad_norm": 20.580381393432617,
+      "learning_rate": 1.8798577325667113e-05,
+      "loss": 0.5637,
+      "step": 97792
+    },
+    {
+      "epoch": 1.204581962706669,
+      "grad_norm": 47.52641677856445,
+      "learning_rate": 1.879543220797091e-05,
+      "loss": 0.6585,
+      "step": 98048
+    },
+    {
+      "epoch": 1.2077271093829185,
+      "grad_norm": 25.38374137878418,
+      "learning_rate": 1.879228709027471e-05,
+      "loss": 0.6784,
+      "step": 98304
+    },
+    {
+      "epoch": 1.210872256059168,
+      "grad_norm": 5.62556266784668,
+      "learning_rate": 1.8789141972578505e-05,
+      "loss": 0.6429,
+      "step": 98560
+    },
+    {
+      "epoch": 1.2140174027354176,
+      "grad_norm": 5.391506195068359,
+      "learning_rate": 1.8785996854882305e-05,
+      "loss": 0.6694,
+      "step": 98816
+    },
+    {
+      "epoch": 1.2171625494116671,
+      "grad_norm": 144.59048461914062,
+      "learning_rate": 1.87828517371861e-05,
+      "loss": 0.7135,
+      "step": 99072
+    },
+    {
+      "epoch": 1.2203076960879167,
+      "grad_norm": 40.11859893798828,
+      "learning_rate": 1.8779706619489905e-05,
+      "loss": 0.7178,
+      "step": 99328
+    },
+    {
+      "epoch": 1.2234528427641662,
+      "grad_norm": 91.49730682373047,
+      "learning_rate": 1.87765615017937e-05,
+      "loss": 0.7157,
+      "step": 99584
+    },
+    {
+      "epoch": 1.2265979894404158,
+      "grad_norm": 7.342621803283691,
+      "learning_rate": 1.87734163840975e-05,
+      "loss": 0.6621,
+      "step": 99840
+    },
+    {
+      "epoch": 1.2297431361166653,
+      "grad_norm": 92.18336486816406,
+      "learning_rate": 1.8770271266401297e-05,
+      "loss": 0.7087,
+      "step": 100096
+    },
+    {
+      "epoch": 1.2328882827929148,
+      "grad_norm": 3.051367998123169,
+      "learning_rate": 1.8767126148705097e-05,
+      "loss": 0.6668,
+      "step": 100352
+    },
+    {
+      "epoch": 1.2360334294691644,
+      "grad_norm": 100.19963836669922,
+      "learning_rate": 1.8763981031008897e-05,
+      "loss": 0.7089,
+      "step": 100608
+    },
+    {
+      "epoch": 1.239178576145414,
+      "grad_norm": 4.8911004066467285,
+      "learning_rate": 1.8760835913312697e-05,
+      "loss": 0.6629,
+      "step": 100864
+    },
+    {
+      "epoch": 1.2423237228216635,
+      "grad_norm": 6.665771484375,
+      "learning_rate": 1.8757690795616493e-05,
+      "loss": 0.6745,
+      "step": 101120
+    },
+    {
+      "epoch": 1.245468869497913,
+      "grad_norm": 39.56829833984375,
+      "learning_rate": 1.8754545677920293e-05,
+      "loss": 0.7356,
+      "step": 101376
+    },
+    {
+      "epoch": 1.2486140161741626,
+      "grad_norm": 47.70098876953125,
+      "learning_rate": 1.8751400560224093e-05,
+      "loss": 0.6372,
+      "step": 101632
+    },
+    {
+      "epoch": 1.251759162850412,
+      "grad_norm": 73.65769958496094,
+      "learning_rate": 1.874825544252789e-05,
+      "loss": 0.6342,
+      "step": 101888
+    },
+    {
+      "epoch": 1.2549043095266614,
+      "grad_norm": 3.158067464828491,
+      "learning_rate": 1.874511032483169e-05,
+      "loss": 0.7064,
+      "step": 102144
+    },
+    {
+      "epoch": 1.2580494562029112,
+      "grad_norm": 31.931184768676758,
+      "learning_rate": 1.8741965207135485e-05,
+      "loss": 0.6746,
+      "step": 102400
+    },
+    {
+      "epoch": 1.2611946028791605,
+      "grad_norm": 32.271175384521484,
+      "learning_rate": 1.8738820089439285e-05,
+      "loss": 0.6787,
+      "step": 102656
+    },
+    {
+      "epoch": 1.2643397495554103,
+      "grad_norm": 41.590904235839844,
+      "learning_rate": 1.8735674971743085e-05,
+      "loss": 0.7497,
+      "step": 102912
+    },
+    {
+      "epoch": 1.2674848962316596,
+      "grad_norm": 0.5061074495315552,
+      "learning_rate": 1.8732529854046885e-05,
+      "loss": 0.7065,
+      "step": 103168
+    },
+    {
+      "epoch": 1.2706300429079094,
+      "grad_norm": 92.25340270996094,
+      "learning_rate": 1.872938473635068e-05,
+      "loss": 0.7947,
+      "step": 103424
+    },
+    {
+      "epoch": 1.2737751895841587,
+      "grad_norm": 200.1078643798828,
+      "learning_rate": 1.872623961865448e-05,
+      "loss": 0.6761,
+      "step": 103680
+    },
+    {
+      "epoch": 1.2769203362604082,
+      "grad_norm": 89.14942932128906,
+      "learning_rate": 1.8723094500958277e-05,
+      "loss": 0.7217,
+      "step": 103936
+    },
+    {
+      "epoch": 1.2800654829366578,
+      "grad_norm": 0.5090438723564148,
+      "learning_rate": 1.871994938326208e-05,
+      "loss": 0.7293,
+      "step": 104192
+    },
+    {
+      "epoch": 1.2832106296129073,
+      "grad_norm": 22.078466415405273,
+      "learning_rate": 1.8716804265565877e-05,
+      "loss": 0.7393,
+      "step": 104448
+    },
+    {
+      "epoch": 1.2863557762891569,
+      "grad_norm": 28.95656394958496,
+      "learning_rate": 1.8713659147869677e-05,
+      "loss": 0.6442,
+      "step": 104704
+    },
+    {
+      "epoch": 1.2895009229654064,
+      "grad_norm": 0.56177818775177,
+      "learning_rate": 1.8710514030173473e-05,
+      "loss": 0.6104,
+      "step": 104960
+    },
+    {
+      "epoch": 1.292646069641656,
+      "grad_norm": 57.65572738647461,
+      "learning_rate": 1.8707368912477273e-05,
+      "loss": 0.5997,
+      "step": 105216
+    },
+    {
+      "epoch": 1.2957912163179055,
+      "grad_norm": 28.302349090576172,
+      "learning_rate": 1.870422379478107e-05,
+      "loss": 0.6666,
+      "step": 105472
+    },
+    {
+      "epoch": 1.298936362994155,
+      "grad_norm": 10.606588363647461,
+      "learning_rate": 1.870107867708487e-05,
+      "loss": 0.6234,
+      "step": 105728
+    },
+    {
+      "epoch": 1.3020815096704046,
+      "grad_norm": 31.015872955322266,
+      "learning_rate": 1.869793355938867e-05,
+      "loss": 0.6927,
+      "step": 105984
+    },
+    {
+      "epoch": 1.3052266563466541,
+      "grad_norm": 53.584529876708984,
+      "learning_rate": 1.869478844169247e-05,
+      "loss": 0.5895,
+      "step": 106240
+    },
+    {
+      "epoch": 1.3083718030229037,
+      "grad_norm": 97.59036254882812,
+      "learning_rate": 1.8691643323996265e-05,
+      "loss": 0.7201,
+      "step": 106496
+    },
+    {
+      "epoch": 1.3115169496991532,
+      "grad_norm": 9.167328834533691,
+      "learning_rate": 1.8688498206300065e-05,
+      "loss": 0.6378,
+      "step": 106752
+    },
+    {
+      "epoch": 1.3146620963754028,
+      "grad_norm": 30.179706573486328,
+      "learning_rate": 1.8685353088603865e-05,
+      "loss": 0.6171,
+      "step": 107008
+    },
+    {
+      "epoch": 1.3178072430516523,
+      "grad_norm": 16.136465072631836,
+      "learning_rate": 1.868220797090766e-05,
+      "loss": 0.6527,
+      "step": 107264
+    },
+    {
+      "epoch": 1.3209523897279019,
+      "grad_norm": 26.04202651977539,
+      "learning_rate": 1.867906285321146e-05,
+      "loss": 0.6966,
+      "step": 107520
+    },
+    {
+      "epoch": 1.3240975364041514,
+      "grad_norm": 206.79830932617188,
+      "learning_rate": 1.867591773551526e-05,
+      "loss": 0.6239,
+      "step": 107776
+    },
+    {
+      "epoch": 1.327242683080401,
+      "grad_norm": 89.86215209960938,
+      "learning_rate": 1.867277261781906e-05,
+      "loss": 0.611,
+      "step": 108032
+    },
+    {
+      "epoch": 1.3303878297566505,
+      "grad_norm": 15.295347213745117,
+      "learning_rate": 1.8669627500122857e-05,
+      "loss": 0.5994,
+      "step": 108288
+    },
+    {
+      "epoch": 1.3335329764329,
+      "grad_norm": 5.113292217254639,
+      "learning_rate": 1.8666482382426657e-05,
+      "loss": 0.749,
+      "step": 108544
+    },
+    {
+      "epoch": 1.3366781231091496,
+      "grad_norm": 60.68025207519531,
+      "learning_rate": 1.8663337264730453e-05,
+      "loss": 0.6878,
+      "step": 108800
+    },
+    {
+      "epoch": 1.339823269785399,
+      "grad_norm": 2.3128774166107178,
+      "learning_rate": 1.8660192147034253e-05,
+      "loss": 0.599,
+      "step": 109056
+    },
+    {
+      "epoch": 1.3429684164616487,
+      "grad_norm": 69.1929702758789,
+      "learning_rate": 1.8657047029338053e-05,
+      "loss": 0.7053,
+      "step": 109312
+    },
+    {
+      "epoch": 1.346113563137898,
+      "grad_norm": 2.1352317333221436,
+      "learning_rate": 1.8653901911641853e-05,
+      "loss": 0.6219,
+      "step": 109568
+    },
+    {
+      "epoch": 1.3492587098141477,
+      "grad_norm": 1.4987752437591553,
+      "learning_rate": 1.865075679394565e-05,
+      "loss": 0.6971,
+      "step": 109824
+    },
+    {
+      "epoch": 1.352403856490397,
+      "grad_norm": 186.3732147216797,
+      "learning_rate": 1.864761167624945e-05,
+      "loss": 0.7012,
+      "step": 110080
+    },
+    {
+      "epoch": 1.3555490031666468,
+      "grad_norm": 13.953139305114746,
+      "learning_rate": 1.8644466558553245e-05,
+      "loss": 0.652,
+      "step": 110336
+    },
+    {
+      "epoch": 1.3586941498428962,
+      "grad_norm": 35.211219787597656,
+      "learning_rate": 1.8641321440857045e-05,
+      "loss": 0.6397,
+      "step": 110592
+    },
+    {
+      "epoch": 1.3618392965191457,
+      "grad_norm": 6.949230194091797,
+      "learning_rate": 1.8638176323160845e-05,
+      "loss": 0.6211,
+      "step": 110848
+    },
+    {
+      "epoch": 1.3649844431953952,
+      "grad_norm": 51.653350830078125,
+      "learning_rate": 1.8635031205464645e-05,
+      "loss": 0.7204,
+      "step": 111104
+    },
+    {
+      "epoch": 1.3681295898716448,
+      "grad_norm": 344.5322265625,
+      "learning_rate": 1.863188608776844e-05,
+      "loss": 0.6209,
+      "step": 111360
+    },
+    {
+      "epoch": 1.3712747365478943,
+      "grad_norm": 181.70057678222656,
+      "learning_rate": 1.862874097007224e-05,
+      "loss": 0.6249,
+      "step": 111616
+    },
+    {
+      "epoch": 1.3744198832241439,
+      "grad_norm": 155.25149536132812,
+      "learning_rate": 1.862559585237604e-05,
+      "loss": 0.7057,
+      "step": 111872
+    },
+    {
+      "epoch": 1.3775650299003934,
+      "grad_norm": 24.911977767944336,
+      "learning_rate": 1.8622450734679837e-05,
+      "loss": 0.7063,
+      "step": 112128
+    },
+    {
+      "epoch": 1.380710176576643,
+      "grad_norm": 23.3475284576416,
+      "learning_rate": 1.8619305616983637e-05,
+      "loss": 0.6866,
+      "step": 112384
+    },
+    {
+      "epoch": 1.3838553232528925,
+      "grad_norm": 30.28309440612793,
+      "learning_rate": 1.8616160499287437e-05,
+      "loss": 0.5971,
+      "step": 112640
+    },
+    {
+      "epoch": 1.387000469929142,
+      "grad_norm": 71.3869400024414,
+      "learning_rate": 1.8613015381591237e-05,
+      "loss": 0.6859,
+      "step": 112896
+    },
+    {
+      "epoch": 1.3901456166053916,
+      "grad_norm": 46.990936279296875,
+      "learning_rate": 1.8609870263895033e-05,
+      "loss": 0.705,
+      "step": 113152
+    },
+    {
+      "epoch": 1.3932907632816411,
+      "grad_norm": 116.8062973022461,
+      "learning_rate": 1.8606725146198833e-05,
+      "loss": 0.6505,
+      "step": 113408
+    },
+    {
+      "epoch": 1.3964359099578907,
+      "grad_norm": 76.77362823486328,
+      "learning_rate": 1.860358002850263e-05,
+      "loss": 0.7222,
+      "step": 113664
+    },
+    {
+      "epoch": 1.3995810566341402,
+      "grad_norm": 67.20590209960938,
+      "learning_rate": 1.860043491080643e-05,
+      "loss": 0.7398,
+      "step": 113920
+    },
+    {
+      "epoch": 1.4027262033103898,
+      "grad_norm": 11.832334518432617,
+      "learning_rate": 1.8597289793110226e-05,
+      "loss": 0.6103,
+      "step": 114176
+    },
+    {
+      "epoch": 1.4058713499866393,
+      "grad_norm": 6.890631198883057,
+      "learning_rate": 1.859414467541403e-05,
+      "loss": 0.7344,
+      "step": 114432
+    },
+    {
+      "epoch": 1.4090164966628889,
+      "grad_norm": 109.55986785888672,
+      "learning_rate": 1.8590999557717825e-05,
+      "loss": 0.6085,
+      "step": 114688
+    },
+    {
+      "epoch": 1.4121616433391384,
+      "grad_norm": 0.32553595304489136,
+      "learning_rate": 1.8587854440021625e-05,
+      "loss": 0.6043,
+      "step": 114944
+    },
+    {
+      "epoch": 1.415306790015388,
+      "grad_norm": 29.583486557006836,
+      "learning_rate": 1.858470932232542e-05,
+      "loss": 0.7446,
+      "step": 115200
+    },
+    {
+      "epoch": 1.4184519366916375,
+      "grad_norm": 9.563398361206055,
+      "learning_rate": 1.858156420462922e-05,
+      "loss": 0.7233,
+      "step": 115456
+    },
+    {
+      "epoch": 1.421597083367887,
+      "grad_norm": 90.88845825195312,
+      "learning_rate": 1.857841908693302e-05,
+      "loss": 0.6637,
+      "step": 115712
+    },
+    {
+      "epoch": 1.4247422300441364,
+      "grad_norm": 15.705517768859863,
+      "learning_rate": 1.857527396923682e-05,
+      "loss": 0.7082,
+      "step": 115968
+    },
+    {
+      "epoch": 1.4278873767203861,
+      "grad_norm": 36.43145751953125,
+      "learning_rate": 1.8572128851540617e-05,
+      "loss": 0.656,
+      "step": 116224
+    },
+    {
+      "epoch": 1.4310325233966354,
+      "grad_norm": 0.27835217118263245,
+      "learning_rate": 1.8568983733844417e-05,
+      "loss": 0.7054,
+      "step": 116480
+    },
+    {
+      "epoch": 1.4341776700728852,
+      "grad_norm": 224.7283935546875,
+      "learning_rate": 1.8565838616148213e-05,
+      "loss": 0.6898,
+      "step": 116736
+    },
+    {
+      "epoch": 1.4373228167491345,
+      "grad_norm": 4.305415153503418,
+      "learning_rate": 1.8562693498452013e-05,
+      "loss": 0.7007,
+      "step": 116992
+    },
+    {
+      "epoch": 1.440467963425384,
+      "grad_norm": 0.7003730535507202,
+      "learning_rate": 1.8559548380755813e-05,
+      "loss": 0.6805,
+      "step": 117248
+    },
+    {
+      "epoch": 1.4436131101016336,
+      "grad_norm": 12.460258483886719,
+      "learning_rate": 1.855640326305961e-05,
+      "loss": 0.5773,
+      "step": 117504
+    },
+    {
+      "epoch": 1.4467582567778832,
+      "grad_norm": 184.58578491210938,
+      "learning_rate": 1.855325814536341e-05,
+      "loss": 0.6466,
+      "step": 117760
+    },
+    {
+      "epoch": 1.4499034034541327,
+      "grad_norm": 56.33019256591797,
+      "learning_rate": 1.855011302766721e-05,
+      "loss": 0.7367,
+      "step": 118016
+    },
+    {
+      "epoch": 1.4530485501303823,
+      "grad_norm": 197.8460693359375,
+      "learning_rate": 1.854696790997101e-05,
+      "loss": 0.7508,
+      "step": 118272
+    },
+    {
+      "epoch": 1.4561936968066318,
+      "grad_norm": 5.9656291007995605,
+      "learning_rate": 1.8543822792274805e-05,
+      "loss": 0.639,
+      "step": 118528
+    },
+    {
+      "epoch": 1.4593388434828813,
+      "grad_norm": 5.656418323516846,
+      "learning_rate": 1.8540677674578605e-05,
+      "loss": 0.6406,
+      "step": 118784
+    },
+    {
+      "epoch": 1.4624839901591309,
+      "grad_norm": 57.2266731262207,
+      "learning_rate": 1.85375325568824e-05,
+      "loss": 0.7175,
+      "step": 119040
+    },
+    {
+      "epoch": 1.4656291368353804,
+      "grad_norm": 20.924455642700195,
+      "learning_rate": 1.8534387439186205e-05,
+      "loss": 0.6301,
+      "step": 119296
+    },
+    {
+      "epoch": 1.46877428351163,
+      "grad_norm": 0.7990367412567139,
+      "learning_rate": 1.853124232149e-05,
+      "loss": 0.718,
+      "step": 119552
+    },
+    {
+      "epoch": 1.4719194301878795,
+      "grad_norm": 72.38453674316406,
+      "learning_rate": 1.85280972037938e-05,
+      "loss": 0.7216,
+      "step": 119808
+    },
+    {
+      "epoch": 1.475064576864129,
+      "grad_norm": 1.432307481765747,
+      "learning_rate": 1.8524952086097597e-05,
+      "loss": 0.6066,
+      "step": 120064
+    },
+    {
+      "epoch": 1.4782097235403786,
+      "grad_norm": 4.171876430511475,
+      "learning_rate": 1.8521806968401397e-05,
+      "loss": 0.6042,
+      "step": 120320
+    },
+    {
+      "epoch": 1.4813548702166282,
+      "grad_norm": 127.643798828125,
+      "learning_rate": 1.8518661850705194e-05,
+      "loss": 0.6857,
+      "step": 120576
+    },
+    {
+      "epoch": 1.4845000168928777,
+      "grad_norm": 66.6846923828125,
+      "learning_rate": 1.8515516733008993e-05,
+      "loss": 0.6693,
+      "step": 120832
+    },
+    {
+      "epoch": 1.4876451635691272,
+      "grad_norm": 38.83477783203125,
+      "learning_rate": 1.8512371615312793e-05,
+      "loss": 0.6432,
+      "step": 121088
+    },
+    {
+      "epoch": 1.4907903102453768,
+      "grad_norm": 92.48795318603516,
+      "learning_rate": 1.8509226497616593e-05,
+      "loss": 0.6332,
+      "step": 121344
+    },
+    {
+      "epoch": 1.4939354569216263,
+      "grad_norm": 5.040313243865967,
+      "learning_rate": 1.850608137992039e-05,
+      "loss": 0.6522,
+      "step": 121600
+    },
+    {
+      "epoch": 1.4970806035978759,
+      "grad_norm": 481.36370849609375,
+      "learning_rate": 1.850293626222419e-05,
+      "loss": 0.7564,
+      "step": 121856
+    },
+    {
+      "epoch": 1.5002257502741254,
+      "grad_norm": 72.99726104736328,
+      "learning_rate": 1.849979114452799e-05,
+      "loss": 0.7585,
+      "step": 122112
+    },
+    {
+      "epoch": 1.5033708969503747,
+      "grad_norm": 39.34186553955078,
+      "learning_rate": 1.8496646026831785e-05,
+      "loss": 0.6765,
+      "step": 122368
+    },
+    {
+      "epoch": 1.5065160436266245,
+      "grad_norm": 144.0253143310547,
+      "learning_rate": 1.8493500909135585e-05,
+      "loss": 0.7271,
+      "step": 122624
+    },
+    {
+      "epoch": 1.5096611903028738,
+      "grad_norm": 185.2099609375,
+      "learning_rate": 1.8490355791439385e-05,
+      "loss": 0.6705,
+      "step": 122880
+    },
+    {
+      "epoch": 1.5128063369791236,
+      "grad_norm": 57.051570892333984,
+      "learning_rate": 1.8487210673743185e-05,
+      "loss": 0.6967,
+      "step": 123136
+    },
+    {
+      "epoch": 1.515951483655373,
+      "grad_norm": 121.9511489868164,
+      "learning_rate": 1.848406555604698e-05,
+      "loss": 0.6602,
+      "step": 123392
+    },
+    {
+      "epoch": 1.5190966303316227,
+      "grad_norm": 49.86616516113281,
+      "learning_rate": 1.848092043835078e-05,
+      "loss": 0.6446,
+      "step": 123648
+    },
+    {
+      "epoch": 1.522241777007872,
+      "grad_norm": 14.083434104919434,
+      "learning_rate": 1.8477775320654578e-05,
+      "loss": 0.7765,
+      "step": 123904
+    },
+    {
+      "epoch": 1.5253869236841218,
+      "grad_norm": 308.0108337402344,
+      "learning_rate": 1.8474630202958377e-05,
+      "loss": 0.6714,
+      "step": 124160
+    },
+    {
+      "epoch": 1.528532070360371,
+      "grad_norm": 211.21383666992188,
+      "learning_rate": 1.8471485085262177e-05,
+      "loss": 0.7136,
+      "step": 124416
+    },
+    {
+      "epoch": 1.5316772170366209,
+      "grad_norm": 13.979375839233398,
+      "learning_rate": 1.8468339967565977e-05,
+      "loss": 0.6498,
+      "step": 124672
+    },
+    {
+      "epoch": 1.5348223637128702,
+      "grad_norm": 1.8429265022277832,
+      "learning_rate": 1.8465194849869773e-05,
+      "loss": 0.7012,
+      "step": 124928
+    },
+    {
+      "epoch": 1.5379675103891197,
+      "grad_norm": 26.227033615112305,
+      "learning_rate": 1.8462049732173573e-05,
+      "loss": 0.6646,
+      "step": 125184
+    },
+    {
+      "epoch": 1.5411126570653693,
+      "grad_norm": 45.74818420410156,
+      "learning_rate": 1.845890461447737e-05,
+      "loss": 0.8054,
+      "step": 125440
+    },
+    {
+      "epoch": 1.5442578037416188,
+      "grad_norm": 123.0312728881836,
+      "learning_rate": 1.845575949678117e-05,
+      "loss": 0.6356,
+      "step": 125696
+    },
+    {
+      "epoch": 1.5474029504178684,
+      "grad_norm": 195.1466522216797,
+      "learning_rate": 1.845261437908497e-05,
+      "loss": 0.6615,
+      "step": 125952
+    },
+    {
+      "epoch": 1.550548097094118,
+      "grad_norm": 13.036552429199219,
+      "learning_rate": 1.844946926138877e-05,
+      "loss": 0.643,
+      "step": 126208
+    },
+    {
+      "epoch": 1.5536932437703674,
+      "grad_norm": 2.844006299972534,
+      "learning_rate": 1.8446324143692565e-05,
+      "loss": 0.6739,
+      "step": 126464
+    },
+    {
+      "epoch": 1.556838390446617,
+      "grad_norm": 3.6921770572662354,
+      "learning_rate": 1.8443179025996365e-05,
+      "loss": 0.709,
+      "step": 126720
+    },
+    {
+      "epoch": 1.5599835371228665,
+      "grad_norm": 0.2538944482803345,
+      "learning_rate": 1.8440033908300165e-05,
+      "loss": 0.6502,
+      "step": 126976
+    },
+    {
+      "epoch": 1.563128683799116,
+      "grad_norm": 5.371120929718018,
+      "learning_rate": 1.843688879060396e-05,
+      "loss": 0.6,
+      "step": 127232
+    },
+    {
+      "epoch": 1.5662738304753656,
+      "grad_norm": 1.362541675567627,
+      "learning_rate": 1.843374367290776e-05,
+      "loss": 0.661,
+      "step": 127488
+    },
+    {
+      "epoch": 1.5694189771516152,
+      "grad_norm": 64.32890319824219,
+      "learning_rate": 1.843059855521156e-05,
+      "loss": 0.641,
+      "step": 127744
+    },
+    {
+      "epoch": 1.5725641238278647,
+      "grad_norm": 111.1507568359375,
+      "learning_rate": 1.8427453437515357e-05,
+      "loss": 0.726,
+      "step": 128000
+    },
+    {
+      "epoch": 1.575709270504114,
+      "grad_norm": 21.845157623291016,
+      "learning_rate": 1.8424308319819157e-05,
+      "loss": 0.6701,
+      "step": 128256
+    },
+    {
+      "epoch": 1.5788544171803638,
+      "grad_norm": 17.954469680786133,
+      "learning_rate": 1.8421163202122957e-05,
+      "loss": 0.5642,
+      "step": 128512
+    },
+    {
+      "epoch": 1.5819995638566131,
+      "grad_norm": 18.45271110534668,
+      "learning_rate": 1.8418018084426753e-05,
+      "loss": 0.7229,
+      "step": 128768
+    },
+    {
+      "epoch": 1.5851447105328629,
+      "grad_norm": 10.551959991455078,
+      "learning_rate": 1.8414872966730553e-05,
+      "loss": 0.6769,
+      "step": 129024
+    },
+    {
+      "epoch": 1.5882898572091122,
+      "grad_norm": 64.96302032470703,
+      "learning_rate": 1.8411727849034353e-05,
+      "loss": 0.654,
+      "step": 129280
+    },
+    {
+      "epoch": 1.591435003885362,
+      "grad_norm": 3.80167818069458,
+      "learning_rate": 1.8408582731338153e-05,
+      "loss": 0.6871,
+      "step": 129536
+    },
+    {
+      "epoch": 1.5945801505616113,
+      "grad_norm": 9.988375663757324,
+      "learning_rate": 1.840543761364195e-05,
+      "loss": 0.6193,
+      "step": 129792
+    },
+    {
+      "epoch": 1.597725297237861,
+      "grad_norm": 163.59457397460938,
+      "learning_rate": 1.840229249594575e-05,
+      "loss": 0.5712,
+      "step": 130048
+    },
+    {
+      "epoch": 1.6008704439141104,
+      "grad_norm": 185.79920959472656,
+      "learning_rate": 1.8399147378249546e-05,
+      "loss": 0.6532,
+      "step": 130304
+    },
+    {
+      "epoch": 1.6040155905903601,
+      "grad_norm": 7.615332126617432,
+      "learning_rate": 1.8396002260553345e-05,
+      "loss": 0.6333,
+      "step": 130560
+    },
+    {
+      "epoch": 1.6071607372666095,
+      "grad_norm": 1.7677505016326904,
+      "learning_rate": 1.8392857142857142e-05,
+      "loss": 0.5989,
+      "step": 130816
+    },
+    {
+      "epoch": 1.6103058839428592,
+      "grad_norm": 127.0537109375,
+      "learning_rate": 1.8389712025160945e-05,
+      "loss": 0.6223,
+      "step": 131072
+    },
+    {
+      "epoch": 1.6134510306191086,
+      "grad_norm": 1.9890645742416382,
+      "learning_rate": 1.838656690746474e-05,
+      "loss": 0.5644,
+      "step": 131328
+    },
+    {
+      "epoch": 1.616596177295358,
+      "grad_norm": 76.5971450805664,
+      "learning_rate": 1.838342178976854e-05,
+      "loss": 0.7171,
+      "step": 131584
+    },
+    {
+      "epoch": 1.6197413239716076,
+      "grad_norm": 0.21291916072368622,
+      "learning_rate": 1.8380276672072338e-05,
+      "loss": 0.7799,
+      "step": 131840
+    },
+    {
+      "epoch": 1.6228864706478572,
+      "grad_norm": 32.992549896240234,
+      "learning_rate": 1.8377131554376137e-05,
+      "loss": 0.6642,
+      "step": 132096
+    },
+    {
+      "epoch": 1.6260316173241067,
+      "grad_norm": 293.4602966308594,
+      "learning_rate": 1.8373986436679937e-05,
+      "loss": 0.5982,
+      "step": 132352
+    },
+    {
+      "epoch": 1.6291767640003563,
+      "grad_norm": 80.05879211425781,
+      "learning_rate": 1.8370841318983734e-05,
+      "loss": 0.6916,
+      "step": 132608
+    },
+    {
+      "epoch": 1.6323219106766058,
+      "grad_norm": 16.697772979736328,
+      "learning_rate": 1.8367696201287533e-05,
+      "loss": 0.6999,
+      "step": 132864
+    },
+    {
+      "epoch": 1.6354670573528554,
+      "grad_norm": 2.0471994876861572,
+      "learning_rate": 1.8364551083591333e-05,
+      "loss": 0.6135,
+      "step": 133120
+    },
+    {
+      "epoch": 1.638612204029105,
+      "grad_norm": 38.79721450805664,
+      "learning_rate": 1.8361405965895133e-05,
+      "loss": 0.6633,
+      "step": 133376
+    },
+    {
+      "epoch": 1.6417573507053544,
+      "grad_norm": 2.6233861446380615,
+      "learning_rate": 1.835826084819893e-05,
+      "loss": 0.6816,
+      "step": 133632
+    },
+    {
+      "epoch": 1.644902497381604,
+      "grad_norm": 9.43591022491455,
+      "learning_rate": 1.835511573050273e-05,
+      "loss": 0.6648,
+      "step": 133888
+    },
+    {
+      "epoch": 1.6480476440578535,
+      "grad_norm": 99.9877700805664,
+      "learning_rate": 1.8351970612806526e-05,
+      "loss": 0.6457,
+      "step": 134144
+    },
+    {
+      "epoch": 1.651192790734103,
+      "grad_norm": 4.154375076293945,
+      "learning_rate": 1.834882549511033e-05,
+      "loss": 0.655,
+      "step": 134400
+    },
+    {
+      "epoch": 1.6543379374103526,
+      "grad_norm": 37.89094924926758,
+      "learning_rate": 1.8345680377414125e-05,
+      "loss": 0.659,
+      "step": 134656
+    },
+    {
+      "epoch": 1.6574830840866022,
+      "grad_norm": 6.969601631164551,
+      "learning_rate": 1.8342535259717925e-05,
+      "loss": 0.5743,
+      "step": 134912
+    },
+    {
+      "epoch": 1.6606282307628515,
+      "grad_norm": 1.3682612180709839,
+      "learning_rate": 1.833939014202172e-05,
+      "loss": 0.6712,
+      "step": 135168
+    },
+    {
+      "epoch": 1.6637733774391013,
+      "grad_norm": 91.62085723876953,
+      "learning_rate": 1.833624502432552e-05,
+      "loss": 0.6581,
+      "step": 135424
+    },
+    {
+      "epoch": 1.6669185241153506,
+      "grad_norm": 0.6061786413192749,
+      "learning_rate": 1.8333099906629318e-05,
+      "loss": 0.6434,
+      "step": 135680
+    },
+    {
+      "epoch": 1.6700636707916003,
+      "grad_norm": 8.49322509765625,
+      "learning_rate": 1.8329954788933117e-05,
+      "loss": 0.6807,
+      "step": 135936
+    },
+    {
+      "epoch": 1.6732088174678497,
+      "grad_norm": 12.335840225219727,
+      "learning_rate": 1.8326809671236917e-05,
+      "loss": 0.6324,
+      "step": 136192
+    },
+    {
+      "epoch": 1.6763539641440994,
+      "grad_norm": 32.167022705078125,
+      "learning_rate": 1.8323664553540717e-05,
+      "loss": 0.685,
+      "step": 136448
+    },
+    {
+      "epoch": 1.6794991108203488,
+      "grad_norm": 14.677338600158691,
+      "learning_rate": 1.8320519435844514e-05,
+      "loss": 0.6583,
+      "step": 136704
+    },
+    {
+      "epoch": 1.6826442574965985,
+      "grad_norm": 34.04314041137695,
+      "learning_rate": 1.8317374318148313e-05,
+      "loss": 0.6069,
+      "step": 136960
+    },
+    {
+      "epoch": 1.6857894041728478,
+      "grad_norm": 51.72367858886719,
+      "learning_rate": 1.8314229200452113e-05,
+      "loss": 0.6301,
+      "step": 137216
+    },
+    {
+      "epoch": 1.6889345508490976,
+      "grad_norm": 43.90719223022461,
+      "learning_rate": 1.831108408275591e-05,
+      "loss": 0.6273,
+      "step": 137472
+    },
+    {
+      "epoch": 1.692079697525347,
+      "grad_norm": 3.7817108631134033,
+      "learning_rate": 1.830793896505971e-05,
+      "loss": 0.713,
+      "step": 137728
+    },
+    {
+      "epoch": 1.6952248442015967,
+      "grad_norm": 41.78171920776367,
+      "learning_rate": 1.830479384736351e-05,
+      "loss": 0.6518,
+      "step": 137984
+    },
+    {
+      "epoch": 1.698369990877846,
+      "grad_norm": 24.83089256286621,
+      "learning_rate": 1.830164872966731e-05,
+      "loss": 0.679,
+      "step": 138240
+    },
+    {
+      "epoch": 1.7015151375540956,
+      "grad_norm": 25.067731857299805,
+      "learning_rate": 1.8298503611971105e-05,
+      "loss": 0.6137,
+      "step": 138496
+    },
+    {
+      "epoch": 1.704660284230345,
+      "grad_norm": 22.014888763427734,
+      "learning_rate": 1.8295358494274905e-05,
+      "loss": 0.6696,
+      "step": 138752
+    },
+    {
+      "epoch": 1.7078054309065946,
+      "grad_norm": 21.678956985473633,
+      "learning_rate": 1.82922133765787e-05,
+      "loss": 0.6865,
+      "step": 139008
+    },
+    {
+      "epoch": 1.7109505775828442,
+      "grad_norm": 0.45430517196655273,
+      "learning_rate": 1.82890682588825e-05,
+      "loss": 0.64,
+      "step": 139264
+    },
+    {
+      "epoch": 1.7140957242590937,
+      "grad_norm": 21.842103958129883,
+      "learning_rate": 1.82859231411863e-05,
+      "loss": 0.6645,
+      "step": 139520
+    },
+    {
+      "epoch": 1.7172408709353433,
+      "grad_norm": 47.032222747802734,
+      "learning_rate": 1.82827780234901e-05,
+      "loss": 0.6316,
+      "step": 139776
+    },
+    {
+      "epoch": 1.7203860176115928,
+      "grad_norm": 0.9306575059890747,
+      "learning_rate": 1.8279632905793897e-05,
+      "loss": 0.6518,
+      "step": 140032
+    },
+    {
+      "epoch": 1.7235311642878424,
+      "grad_norm": 54.06832504272461,
+      "learning_rate": 1.8276487788097697e-05,
+      "loss": 0.6582,
+      "step": 140288
+    },
+    {
+      "epoch": 1.726676310964092,
+      "grad_norm": 15.14962387084961,
+      "learning_rate": 1.8273342670401494e-05,
+      "loss": 0.7038,
+      "step": 140544
+    },
+    {
+      "epoch": 1.7298214576403415,
+      "grad_norm": 18.525476455688477,
+      "learning_rate": 1.8270197552705293e-05,
+      "loss": 0.7372,
+      "step": 140800
+    },
+    {
+      "epoch": 1.732966604316591,
+      "grad_norm": 1.7285302877426147,
+      "learning_rate": 1.8267052435009093e-05,
+      "loss": 0.6251,
+      "step": 141056
+    },
+    {
+      "epoch": 1.7361117509928405,
+      "grad_norm": 23.482906341552734,
+      "learning_rate": 1.8263907317312893e-05,
+      "loss": 0.5649,
+      "step": 141312
+    },
+    {
+      "epoch": 1.7392568976690899,
+      "grad_norm": 12.098392486572266,
+      "learning_rate": 1.826076219961669e-05,
+      "loss": 0.6954,
+      "step": 141568
+    },
+    {
+      "epoch": 1.7424020443453396,
+      "grad_norm": 536.5809936523438,
+      "learning_rate": 1.825761708192049e-05,
+      "loss": 0.668,
+      "step": 141824
+    },
+    {
+      "epoch": 1.745547191021589,
+      "grad_norm": 16.067285537719727,
+      "learning_rate": 1.8254471964224286e-05,
+      "loss": 0.6549,
+      "step": 142080
+    },
+    {
+      "epoch": 1.7486923376978387,
+      "grad_norm": 10.559584617614746,
+      "learning_rate": 1.8251326846528085e-05,
+      "loss": 0.6688,
+      "step": 142336
+    },
+    {
+      "epoch": 1.751837484374088,
+      "grad_norm": 70.58854675292969,
+      "learning_rate": 1.8248181728831885e-05,
+      "loss": 0.6469,
+      "step": 142592
+    },
+    {
+      "epoch": 1.7549826310503378,
+      "grad_norm": 5.755173206329346,
+      "learning_rate": 1.8245036611135685e-05,
+      "loss": 0.6677,
+      "step": 142848
+    },
+    {
+      "epoch": 1.7581277777265871,
+      "grad_norm": 11.262292861938477,
+      "learning_rate": 1.824189149343948e-05,
+      "loss": 0.6265,
+      "step": 143104
+    },
+    {
+      "epoch": 1.761272924402837,
+      "grad_norm": 13.489733695983887,
+      "learning_rate": 1.823874637574328e-05,
+      "loss": 0.7508,
+      "step": 143360
+    },
+    {
+      "epoch": 1.7644180710790862,
+      "grad_norm": 26.479047775268555,
+      "learning_rate": 1.823560125804708e-05,
+      "loss": 0.5799,
+      "step": 143616
+    },
+    {
+      "epoch": 1.767563217755336,
+      "grad_norm": 106.04133605957031,
+      "learning_rate": 1.8232456140350878e-05,
+      "loss": 0.6373,
+      "step": 143872
+    },
+    {
+      "epoch": 1.7707083644315853,
+      "grad_norm": 3.0954108238220215,
+      "learning_rate": 1.8229311022654677e-05,
+      "loss": 0.6193,
+      "step": 144128
+    },
+    {
+      "epoch": 1.773853511107835,
+      "grad_norm": 61.435184478759766,
+      "learning_rate": 1.8226165904958477e-05,
+      "loss": 0.6909,
+      "step": 144384
+    },
+    {
+      "epoch": 1.7769986577840844,
+      "grad_norm": 39.79731369018555,
+      "learning_rate": 1.8223020787262277e-05,
+      "loss": 0.7179,
+      "step": 144640
+    },
+    {
+      "epoch": 1.780143804460334,
+      "grad_norm": 7.779622554779053,
+      "learning_rate": 1.8219875669566073e-05,
+      "loss": 0.5834,
+      "step": 144896
+    },
+    {
+      "epoch": 1.7832889511365835,
+      "grad_norm": 0.26465222239494324,
+      "learning_rate": 1.8216730551869873e-05,
+      "loss": 0.573,
+      "step": 145152
+    },
+    {
+      "epoch": 1.786434097812833,
+      "grad_norm": 9.721681594848633,
+      "learning_rate": 1.821358543417367e-05,
+      "loss": 0.6757,
+      "step": 145408
+    },
+    {
+      "epoch": 1.7895792444890826,
+      "grad_norm": 16.63478660583496,
+      "learning_rate": 1.821044031647747e-05,
+      "loss": 0.6641,
+      "step": 145664
+    },
+    {
+      "epoch": 1.792724391165332,
+      "grad_norm": 9.504838943481445,
+      "learning_rate": 1.8207295198781266e-05,
+      "loss": 0.7078,
+      "step": 145920
+    },
+    {
+      "epoch": 1.7958695378415817,
+      "grad_norm": 5.8867950439453125,
+      "learning_rate": 1.820415008108507e-05,
+      "loss": 0.6676,
+      "step": 146176
+    },
+    {
+      "epoch": 1.7990146845178312,
+      "grad_norm": 197.24130249023438,
+      "learning_rate": 1.8201004963388865e-05,
+      "loss": 0.6611,
+      "step": 146432
+    },
+    {
+      "epoch": 1.8021598311940807,
+      "grad_norm": 3.12298846244812,
+      "learning_rate": 1.8197859845692665e-05,
+      "loss": 0.6231,
+      "step": 146688
+    },
+    {
+      "epoch": 1.8053049778703303,
+      "grad_norm": 3.116199016571045,
+      "learning_rate": 1.819471472799646e-05,
+      "loss": 0.7698,
+      "step": 146944
+    },
+    {
+      "epoch": 1.8084501245465798,
+      "grad_norm": 11.110424041748047,
+      "learning_rate": 1.819156961030026e-05,
+      "loss": 0.6128,
+      "step": 147200
+    },
+    {
+      "epoch": 1.8115952712228294,
+      "grad_norm": 1.2738146781921387,
+      "learning_rate": 1.818842449260406e-05,
+      "loss": 0.7263,
+      "step": 147456
+    },
+    {
+      "epoch": 1.814740417899079,
+      "grad_norm": 26.259117126464844,
+      "learning_rate": 1.818527937490786e-05,
+      "loss": 0.6725,
+      "step": 147712
+    },
+    {
+      "epoch": 1.8178855645753285,
+      "grad_norm": 18.465112686157227,
+      "learning_rate": 1.8182134257211657e-05,
+      "loss": 0.6415,
+      "step": 147968
+    },
+    {
+      "epoch": 1.821030711251578,
+      "grad_norm": 137.1578369140625,
+      "learning_rate": 1.8178989139515457e-05,
+      "loss": 0.6422,
+      "step": 148224
+    },
+    {
+      "epoch": 1.8241758579278273,
+      "grad_norm": 1.390212059020996,
+      "learning_rate": 1.8175844021819257e-05,
+      "loss": 0.6812,
+      "step": 148480
+    },
+    {
+      "epoch": 1.827321004604077,
+      "grad_norm": 76.83866882324219,
+      "learning_rate": 1.8172698904123053e-05,
+      "loss": 0.6463,
+      "step": 148736
+    },
+    {
+      "epoch": 1.8304661512803264,
+      "grad_norm": 3.9307312965393066,
+      "learning_rate": 1.8169553786426853e-05,
+      "loss": 0.6603,
+      "step": 148992
+    },
+    {
+      "epoch": 1.8336112979565762,
+      "grad_norm": 17.38741683959961,
+      "learning_rate": 1.816640866873065e-05,
+      "loss": 0.5558,
+      "step": 149248
+    },
+    {
+      "epoch": 1.8367564446328255,
+      "grad_norm": 0.33097225427627563,
+      "learning_rate": 1.8163263551034453e-05,
+      "loss": 0.6592,
+      "step": 149504
+    },
+    {
+      "epoch": 1.8399015913090753,
+      "grad_norm": 57.788509368896484,
+      "learning_rate": 1.816011843333825e-05,
+      "loss": 0.7214,
+      "step": 149760
+    },
+    {
+      "epoch": 1.8430467379853246,
+      "grad_norm": 208.28578186035156,
+      "learning_rate": 1.815697331564205e-05,
+      "loss": 0.6611,
+      "step": 150016
+    },
+    {
+      "epoch": 1.8461918846615744,
+      "grad_norm": 55.99864959716797,
+      "learning_rate": 1.8153828197945846e-05,
+      "loss": 0.624,
+      "step": 150272
+    },
+    {
+      "epoch": 1.8493370313378237,
+      "grad_norm": 34.706600189208984,
+      "learning_rate": 1.8150683080249645e-05,
+      "loss": 0.6383,
+      "step": 150528
+    },
+    {
+      "epoch": 1.8524821780140734,
+      "grad_norm": 65.20597839355469,
+      "learning_rate": 1.8147537962553442e-05,
+      "loss": 0.6228,
+      "step": 150784
+    },
+    {
+      "epoch": 1.8556273246903228,
+      "grad_norm": 3.494520425796509,
+      "learning_rate": 1.8144392844857245e-05,
+      "loss": 0.6557,
+      "step": 151040
+    },
+    {
+      "epoch": 1.8587724713665725,
+      "grad_norm": 22.704248428344727,
+      "learning_rate": 1.814124772716104e-05,
+      "loss": 0.7322,
+      "step": 151296
+    },
+    {
+      "epoch": 1.8619176180428219,
+      "grad_norm": 0.8666673898696899,
+      "learning_rate": 1.813810260946484e-05,
+      "loss": 0.6282,
+      "step": 151552
+    },
+    {
+      "epoch": 1.8650627647190714,
+      "grad_norm": 1.25579035282135,
+      "learning_rate": 1.8134957491768638e-05,
+      "loss": 0.711,
+      "step": 151808
+    },
+    {
+      "epoch": 1.868207911395321,
+      "grad_norm": 18.007038116455078,
+      "learning_rate": 1.8131812374072437e-05,
+      "loss": 0.6735,
+      "step": 152064
+    },
+    {
+      "epoch": 1.8713530580715705,
+      "grad_norm": 62.08824157714844,
+      "learning_rate": 1.8128667256376237e-05,
+      "loss": 0.6247,
+      "step": 152320
+    },
+    {
+      "epoch": 1.87449820474782,
+      "grad_norm": 22.543054580688477,
+      "learning_rate": 1.8125522138680034e-05,
+      "loss": 0.5332,
+      "step": 152576
+    },
+    {
+      "epoch": 1.8776433514240696,
+      "grad_norm": 22.97509765625,
+      "learning_rate": 1.8122377020983833e-05,
+      "loss": 0.6311,
+      "step": 152832
+    },
+    {
+      "epoch": 1.8807884981003191,
+      "grad_norm": 22.249046325683594,
+      "learning_rate": 1.8119231903287633e-05,
+      "loss": 0.6897,
+      "step": 153088
+    },
+    {
+      "epoch": 1.8839336447765687,
+      "grad_norm": 4.590738296508789,
+      "learning_rate": 1.811608678559143e-05,
+      "loss": 0.6492,
+      "step": 153344
+    },
+    {
+      "epoch": 1.8870787914528182,
+      "grad_norm": 4.006165027618408,
+      "learning_rate": 1.811294166789523e-05,
+      "loss": 0.6326,
+      "step": 153600
+    },
+    {
+      "epoch": 1.8902239381290677,
+      "grad_norm": 24.309553146362305,
+      "learning_rate": 1.810979655019903e-05,
+      "loss": 0.6,
+      "step": 153856
+    },
+    {
+      "epoch": 1.8933690848053173,
+      "grad_norm": 48.84052276611328,
+      "learning_rate": 1.8106651432502826e-05,
+      "loss": 0.576,
+      "step": 154112
+    },
+    {
+      "epoch": 1.8965142314815668,
+      "grad_norm": 26.35711097717285,
+      "learning_rate": 1.8103506314806625e-05,
+      "loss": 0.6663,
+      "step": 154368
+    },
+    {
+      "epoch": 1.8996593781578164,
+      "grad_norm": 45.187095642089844,
+      "learning_rate": 1.8100361197110425e-05,
+      "loss": 0.7006,
+      "step": 154624
+    },
+    {
+      "epoch": 1.902804524834066,
+      "grad_norm": 3.7311763763427734,
+      "learning_rate": 1.8097216079414225e-05,
+      "loss": 0.6554,
+      "step": 154880
+    },
+    {
+      "epoch": 1.9059496715103155,
+      "grad_norm": 1.1564083099365234,
+      "learning_rate": 1.809407096171802e-05,
+      "loss": 0.6048,
+      "step": 155136
+    },
+    {
+      "epoch": 1.9090948181865648,
+      "grad_norm": 0.20646964013576508,
+      "learning_rate": 1.809092584402182e-05,
+      "loss": 0.6374,
+      "step": 155392
+    },
+    {
+      "epoch": 1.9122399648628146,
+      "grad_norm": 0.43324726819992065,
+      "learning_rate": 1.8087780726325618e-05,
+      "loss": 0.5852,
+      "step": 155648
+    },
+    {
+      "epoch": 1.9153851115390639,
+      "grad_norm": 23.553667068481445,
+      "learning_rate": 1.8084635608629418e-05,
+      "loss": 0.6041,
+      "step": 155904
+    },
+    {
+      "epoch": 1.9185302582153136,
+      "grad_norm": 33.44930648803711,
+      "learning_rate": 1.8081490490933217e-05,
+      "loss": 0.6703,
+      "step": 156160
+    },
+    {
+      "epoch": 1.921675404891563,
+      "grad_norm": 72.30155181884766,
+      "learning_rate": 1.8078345373237017e-05,
+      "loss": 0.6869,
+      "step": 156416
+    },
+    {
+      "epoch": 1.9248205515678127,
+      "grad_norm": 28.453449249267578,
+      "learning_rate": 1.8075200255540814e-05,
+      "loss": 0.6979,
+      "step": 156672
+    },
+    {
+      "epoch": 1.927965698244062,
+      "grad_norm": 39.39230728149414,
+      "learning_rate": 1.8072055137844613e-05,
+      "loss": 0.7512,
+      "step": 156928
+    },
+    {
+      "epoch": 1.9311108449203118,
+      "grad_norm": 64.26713562011719,
+      "learning_rate": 1.806891002014841e-05,
+      "loss": 0.6477,
+      "step": 157184
+    },
+    {
+      "epoch": 1.9342559915965611,
+      "grad_norm": 2.1566553115844727,
+      "learning_rate": 1.806576490245221e-05,
+      "loss": 0.7127,
+      "step": 157440
+    },
+    {
+      "epoch": 1.937401138272811,
+      "grad_norm": 64.05670166015625,
+      "learning_rate": 1.806261978475601e-05,
+      "loss": 0.6254,
+      "step": 157696
+    },
+    {
+      "epoch": 1.9405462849490602,
+      "grad_norm": 2.9898900985717773,
+      "learning_rate": 1.805947466705981e-05,
+      "loss": 0.7016,
+      "step": 157952
+    },
+    {
+      "epoch": 1.94369143162531,
+      "grad_norm": 0.575606644153595,
+      "learning_rate": 1.8056329549363606e-05,
+      "loss": 0.6346,
+      "step": 158208
+    },
+    {
+      "epoch": 1.9468365783015593,
+      "grad_norm": 2.89739990234375,
+      "learning_rate": 1.8053184431667405e-05,
+      "loss": 0.5779,
+      "step": 158464
+    },
+    {
+      "epoch": 1.9499817249778089,
+      "grad_norm": 49.22926712036133,
+      "learning_rate": 1.8050039313971205e-05,
+      "loss": 0.5976,
+      "step": 158720
+    },
+    {
+      "epoch": 1.9531268716540584,
+      "grad_norm": 1.9407929182052612,
+      "learning_rate": 1.8046894196275e-05,
+      "loss": 0.7562,
+      "step": 158976
+    },
+    {
+      "epoch": 1.956272018330308,
+      "grad_norm": 46.429813385009766,
+      "learning_rate": 1.80437490785788e-05,
+      "loss": 0.6995,
+      "step": 159232
+    },
+    {
+      "epoch": 1.9594171650065575,
+      "grad_norm": 98.60050201416016,
+      "learning_rate": 1.80406039608826e-05,
+      "loss": 0.6406,
+      "step": 159488
+    },
+    {
+      "epoch": 1.962562311682807,
+      "grad_norm": 34.39356231689453,
+      "learning_rate": 1.80374588431864e-05,
+      "loss": 0.644,
+      "step": 159744
+    },
+    {
+      "epoch": 1.9657074583590566,
+      "grad_norm": 27.551597595214844,
+      "learning_rate": 1.8034313725490197e-05,
+      "loss": 0.723,
+      "step": 160000
+    },
+    {
+      "epoch": 1.9688526050353061,
+      "grad_norm": 50.80730438232422,
+      "learning_rate": 1.8031168607793997e-05,
+      "loss": 0.586,
+      "step": 160256
+    },
+    {
+      "epoch": 1.9719977517115557,
+      "grad_norm": 60.13485336303711,
+      "learning_rate": 1.8028023490097794e-05,
+      "loss": 0.6527,
+      "step": 160512
+    },
+    {
+      "epoch": 1.9751428983878052,
+      "grad_norm": 3.8352346420288086,
+      "learning_rate": 1.8024878372401593e-05,
+      "loss": 0.6554,
+      "step": 160768
+    },
+    {
+      "epoch": 1.9782880450640548,
+      "grad_norm": 31.221696853637695,
+      "learning_rate": 1.802173325470539e-05,
+      "loss": 0.5922,
+      "step": 161024
+    },
+    {
+      "epoch": 1.9814331917403043,
+      "grad_norm": 26.26588249206543,
+      "learning_rate": 1.8018588137009193e-05,
+      "loss": 0.5952,
+      "step": 161280
+    },
+    {
+      "epoch": 1.9845783384165538,
+      "grad_norm": 59.96664047241211,
+      "learning_rate": 1.801544301931299e-05,
+      "loss": 0.6315,
+      "step": 161536
+    },
+    {
+      "epoch": 1.9877234850928032,
+      "grad_norm": 0.5425397157669067,
+      "learning_rate": 1.801229790161679e-05,
+      "loss": 0.6765,
+      "step": 161792
+    },
+    {
+      "epoch": 1.990868631769053,
+      "grad_norm": 18.745838165283203,
+      "learning_rate": 1.8009152783920586e-05,
+      "loss": 0.6347,
+      "step": 162048
+    },
+    {
+      "epoch": 1.9940137784453023,
+      "grad_norm": 13.62142562866211,
+      "learning_rate": 1.8006007666224386e-05,
+      "loss": 0.6373,
+      "step": 162304
+    },
+    {
+      "epoch": 1.997158925121552,
+      "grad_norm": 2.6074626445770264,
+      "learning_rate": 1.8002862548528185e-05,
+      "loss": 0.6877,
+      "step": 162560
+    },
+    {
+      "epoch": 2.0,
+      "eval_AGE": {
+        "f1": 0.9523482621218679,
+        "number": 32192,
+        "precision": 0.9164727847192302,
+        "recall": 0.9911468687872763
+      },
+      "eval_BUILDINGNUM": {
+        "f1": 0.9711091036151972,
+        "number": 28889,
+        "precision": 0.95658159838818,
+        "recall": 0.9860846689051196
+      },
+      "eval_CITY": {
+        "f1": 0.9496872259573224,
+        "number": 41669,
+        "precision": 0.9260078438526086,
+        "recall": 0.9746094218723751
+      },
+      "eval_DATE": {
+        "f1": 0.9967253957584317,
+        "number": 80940,
+        "precision": 0.995025671965229,
+        "recall": 0.99843093649617
+      },
+      "eval_EMAIL": {
+        "f1": 0.9949708737864078,
+        "number": 51445,
+        "precision": 0.9939094171273397,
+        "recall": 0.9960346000583147
+      },
+      "eval_GENDER": {
+        "f1": 0.96667333009681,
+        "number": 17457,
+        "precision": 0.9638382687927107,
+        "recall": 0.9695251188634931
+      },
+      "eval_GIVENNAME": {
+        "f1": 0.9117306143297775,
+        "number": 100350,
+        "precision": 0.8989094853468147,
+        "recall": 0.9249227703039362
+      },
+      "eval_IDCARDNUM": {
+        "f1": 0.893014570169153,
+        "number": 22068,
+        "precision": 0.8573274963518865,
+        "recall": 0.9318017038245423
+      },
+      "eval_PASSPORTNUM": {
+        "f1": 0.9147639062440834,
+        "number": 12959,
+        "precision": 0.8980669144981412,
+        "recall": 0.9320935257350104
+      },
+      "eval_SEX": {
+        "f1": 0.9651221909636307,
+        "number": 14570,
+        "precision": 0.9533918167816245,
+        "recall": 0.9771448181194234
+      },
+      "eval_SOCIALNUM": {
+        "f1": 0.8993399339933995,
+        "number": 15327,
+        "precision": 0.8458354888773927,
+        "recall": 0.9600704638872578
+      },
+      "eval_STREET": {
+        "f1": 0.9731305914323227,
+        "number": 32803,
+        "precision": 0.974678124713294,
+        "recall": 0.9715879645154407
+      },
+      "eval_SURNAME": {
+        "f1": 0.9018103569256678,
+        "number": 85192,
+        "precision": 0.8984899328859061,
+        "recall": 0.9051554136538642
+      },
+      "eval_TELEPHONENUM": {
+        "f1": 0.9892235870252526,
+        "number": 36941,
+        "precision": 0.9833101529902643,
+        "recall": 0.9952085758371457
+      },
+      "eval_ZIPCODE": {
+        "f1": 0.9735227573836652,
+        "number": 24346,
+        "precision": 0.9662553105401578,
+        "recall": 0.9809003532407787
+      },
+      "eval_overall_accuracy": 0.9910379795542931,
+      "eval_overall_f1": 0.9491292816656416,
+      "eval_overall_precision": 0.9372889547832561,
+      "eval_overall_recall": 0.9612725823413961,
+      "eval_runtime": 3713.0482,
+      "eval_samples_per_second": 87.686,
+      "eval_steps_per_second": 43.843,
+      "step": 162792
+    },
+    {
+      "epoch": 2.0002948575008985,
+      "grad_norm": 2.4750266075134277,
+      "learning_rate": 1.7999717430831985e-05,
+      "loss": 0.766,
+      "step": 162816
+    },
+    {
+      "epoch": 2.003440004177148,
+      "grad_norm": 3.1917426586151123,
+      "learning_rate": 1.799657231313578e-05,
+      "loss": 0.5308,
+      "step": 163072
+    },
+    {
+      "epoch": 2.0065851508533976,
+      "grad_norm": 76.5118637084961,
+      "learning_rate": 1.799342719543958e-05,
+      "loss": 0.5507,
+      "step": 163328
+    },
+    {
+      "epoch": 2.009730297529647,
+      "grad_norm": 163.30284118652344,
+      "learning_rate": 1.799028207774338e-05,
+      "loss": 0.4896,
+      "step": 163584
+    },
+    {
+      "epoch": 2.0128754442058967,
+      "grad_norm": 12.261972427368164,
+      "learning_rate": 1.7987136960047178e-05,
+      "loss": 0.5341,
+      "step": 163840
+    },
+    {
+      "epoch": 2.016020590882146,
+      "grad_norm": 17.510791778564453,
+      "learning_rate": 1.7983991842350977e-05,
+      "loss": 0.5586,
+      "step": 164096
+    },
+    {
+      "epoch": 2.019165737558396,
+      "grad_norm": 0.579840898513794,
+      "learning_rate": 1.7980846724654774e-05,
+      "loss": 0.5127,
+      "step": 164352
+    },
+    {
+      "epoch": 2.022310884234645,
+      "grad_norm": 32.422725677490234,
+      "learning_rate": 1.7977701606958574e-05,
+      "loss": 0.5947,
+      "step": 164608
+    },
+    {
+      "epoch": 2.025456030910895,
+      "grad_norm": 0.14638854563236237,
+      "learning_rate": 1.7974556489262373e-05,
+      "loss": 0.6076,
+      "step": 164864
+    },
+    {
+      "epoch": 2.028601177587144,
+      "grad_norm": 4.373912811279297,
+      "learning_rate": 1.7971411371566173e-05,
+      "loss": 0.6063,
+      "step": 165120
+    },
+    {
+      "epoch": 2.031746324263394,
+      "grad_norm": 0.7967185974121094,
+      "learning_rate": 1.796826625386997e-05,
+      "loss": 0.6473,
+      "step": 165376
+    },
+    {
+      "epoch": 2.0348914709396433,
+      "grad_norm": 88.95084381103516,
+      "learning_rate": 1.796512113617377e-05,
+      "loss": 0.5016,
+      "step": 165632
+    },
+    {
+      "epoch": 2.0380366176158926,
+      "grad_norm": 3.4898133277893066,
+      "learning_rate": 1.7961976018477566e-05,
+      "loss": 0.5873,
+      "step": 165888
+    },
+    {
+      "epoch": 2.0411817642921424,
+      "grad_norm": 35.405677795410156,
+      "learning_rate": 1.795883090078137e-05,
+      "loss": 0.6071,
+      "step": 166144
+    },
+    {
+      "epoch": 2.0443269109683917,
+      "grad_norm": 29.587814331054688,
+      "learning_rate": 1.7955685783085165e-05,
+      "loss": 0.5925,
+      "step": 166400
+    },
+    {
+      "epoch": 2.0474720576446415,
+      "grad_norm": 10.698874473571777,
+      "learning_rate": 1.7952540665388965e-05,
+      "loss": 0.5915,
+      "step": 166656
+    },
+    {
+      "epoch": 2.050617204320891,
+      "grad_norm": 0.34661704301834106,
+      "learning_rate": 1.794939554769276e-05,
+      "loss": 0.4837,
+      "step": 166912
+    },
+    {
+      "epoch": 2.0537623509971406,
+      "grad_norm": 36.818572998046875,
+      "learning_rate": 1.794625042999656e-05,
+      "loss": 0.4975,
+      "step": 167168
+    },
+    {
+      "epoch": 2.05690749767339,
+      "grad_norm": 42.23927307128906,
+      "learning_rate": 1.7943105312300358e-05,
+      "loss": 0.5837,
+      "step": 167424
+    },
+    {
+      "epoch": 2.0600526443496396,
+      "grad_norm": 0.2654969096183777,
+      "learning_rate": 1.7939960194604158e-05,
+      "loss": 0.5952,
+      "step": 167680
+    },
+    {
+      "epoch": 2.063197791025889,
+      "grad_norm": 0.4814351201057434,
+      "learning_rate": 1.7936815076907957e-05,
+      "loss": 0.6356,
+      "step": 167936
+    },
+    {
+      "epoch": 2.0663429377021387,
+      "grad_norm": 31.562015533447266,
+      "learning_rate": 1.7933669959211757e-05,
+      "loss": 0.5701,
+      "step": 168192
+    },
+    {
+      "epoch": 2.069488084378388,
+      "grad_norm": 12.579757690429688,
+      "learning_rate": 1.7930524841515554e-05,
+      "loss": 0.5631,
+      "step": 168448
+    },
+    {
+      "epoch": 2.072633231054638,
+      "grad_norm": 9.925857543945312,
+      "learning_rate": 1.7927379723819354e-05,
+      "loss": 0.5403,
+      "step": 168704
+    },
+    {
+      "epoch": 2.075778377730887,
+      "grad_norm": 0.050978172570466995,
+      "learning_rate": 1.7924234606123153e-05,
+      "loss": 0.5645,
+      "step": 168960
+    },
+    {
+      "epoch": 2.078923524407137,
+      "grad_norm": 0.19265970587730408,
+      "learning_rate": 1.792108948842695e-05,
+      "loss": 0.4889,
+      "step": 169216
+    },
+    {
+      "epoch": 2.0820686710833862,
+      "grad_norm": 3.266355514526367,
+      "learning_rate": 1.791794437073075e-05,
+      "loss": 0.581,
+      "step": 169472
+    },
+    {
+      "epoch": 2.085213817759636,
+      "grad_norm": 72.87947845458984,
+      "learning_rate": 1.791479925303455e-05,
+      "loss": 0.558,
+      "step": 169728
+    },
+    {
+      "epoch": 2.0883589644358853,
+      "grad_norm": 74.64995574951172,
+      "learning_rate": 1.791165413533835e-05,
+      "loss": 0.5541,
+      "step": 169984
+    },
+    {
+      "epoch": 2.091504111112135,
+      "grad_norm": 1.039330244064331,
+      "learning_rate": 1.7908509017642146e-05,
+      "loss": 0.5499,
+      "step": 170240
+    },
+    {
+      "epoch": 2.0946492577883844,
+      "grad_norm": 3.5702216625213623,
+      "learning_rate": 1.7905363899945945e-05,
+      "loss": 0.5736,
+      "step": 170496
+    },
+    {
+      "epoch": 2.097794404464634,
+      "grad_norm": 8.600748062133789,
+      "learning_rate": 1.7902218782249742e-05,
+      "loss": 0.6102,
+      "step": 170752
+    },
+    {
+      "epoch": 2.1009395511408835,
+      "grad_norm": 215.6809539794922,
+      "learning_rate": 1.789907366455354e-05,
+      "loss": 0.484,
+      "step": 171008
+    },
+    {
+      "epoch": 2.1040846978171333,
+      "grad_norm": 41.94133377075195,
+      "learning_rate": 1.789592854685734e-05,
+      "loss": 0.5928,
+      "step": 171264
+    },
+    {
+      "epoch": 2.1072298444933826,
+      "grad_norm": 42.71377182006836,
+      "learning_rate": 1.789278342916114e-05,
+      "loss": 0.5582,
+      "step": 171520
+    },
+    {
+      "epoch": 2.1103749911696323,
+      "grad_norm": 162.4323272705078,
+      "learning_rate": 1.7889638311464938e-05,
+      "loss": 0.5142,
+      "step": 171776
+    },
+    {
+      "epoch": 2.1135201378458817,
+      "grad_norm": 73.68289947509766,
+      "learning_rate": 1.7886493193768737e-05,
+      "loss": 0.614,
+      "step": 172032
+    },
+    {
+      "epoch": 2.1166652845221314,
+      "grad_norm": 59.87325668334961,
+      "learning_rate": 1.7883348076072534e-05,
+      "loss": 0.6868,
+      "step": 172288
+    },
+    {
+      "epoch": 2.1198104311983808,
+      "grad_norm": 7.3754682540893555,
+      "learning_rate": 1.7880202958376334e-05,
+      "loss": 0.5943,
+      "step": 172544
+    },
+    {
+      "epoch": 2.12295557787463,
+      "grad_norm": 14.087455749511719,
+      "learning_rate": 1.7877057840680133e-05,
+      "loss": 0.6176,
+      "step": 172800
+    },
+    {
+      "epoch": 2.12610072455088,
+      "grad_norm": 37.316001892089844,
+      "learning_rate": 1.7873912722983933e-05,
+      "loss": 0.6193,
+      "step": 173056
+    },
+    {
+      "epoch": 2.129245871227129,
+      "grad_norm": 9.935317039489746,
+      "learning_rate": 1.787076760528773e-05,
+      "loss": 0.5299,
+      "step": 173312
+    },
+    {
+      "epoch": 2.132391017903379,
+      "grad_norm": 74.93743133544922,
+      "learning_rate": 1.786762248759153e-05,
+      "loss": 0.6045,
+      "step": 173568
+    },
+    {
+      "epoch": 2.1355361645796282,
+      "grad_norm": 57.48478317260742,
+      "learning_rate": 1.786447736989533e-05,
+      "loss": 0.5091,
+      "step": 173824
+    },
+    {
+      "epoch": 2.138681311255878,
+      "grad_norm": 0.18920351564884186,
+      "learning_rate": 1.7861332252199126e-05,
+      "loss": 0.5643,
+      "step": 174080
+    },
+    {
+      "epoch": 2.1418264579321273,
+      "grad_norm": 74.21656799316406,
+      "learning_rate": 1.7858187134502925e-05,
+      "loss": 0.6741,
+      "step": 174336
+    },
+    {
+      "epoch": 2.144971604608377,
+      "grad_norm": 63.73550033569336,
+      "learning_rate": 1.7855042016806725e-05,
+      "loss": 0.5041,
+      "step": 174592
+    },
+    {
+      "epoch": 2.1481167512846264,
+      "grad_norm": 10.951518058776855,
+      "learning_rate": 1.7851896899110525e-05,
+      "loss": 0.5887,
+      "step": 174848
+    },
+    {
+      "epoch": 2.151261897960876,
+      "grad_norm": 10.460653305053711,
+      "learning_rate": 1.784875178141432e-05,
+      "loss": 0.5783,
+      "step": 175104
+    },
+    {
+      "epoch": 2.1544070446371255,
+      "grad_norm": 114.88678741455078,
+      "learning_rate": 1.784560666371812e-05,
+      "loss": 0.5208,
+      "step": 175360
+    },
+    {
+      "epoch": 2.1575521913133753,
+      "grad_norm": 3.5670249462127686,
+      "learning_rate": 1.7842461546021918e-05,
+      "loss": 0.689,
+      "step": 175616
+    },
+    {
+      "epoch": 2.1606973379896246,
+      "grad_norm": 1.8214390277862549,
+      "learning_rate": 1.7839316428325718e-05,
+      "loss": 0.6452,
+      "step": 175872
+    },
+    {
+      "epoch": 2.1638424846658744,
+      "grad_norm": 118.2511215209961,
+      "learning_rate": 1.7836171310629514e-05,
+      "loss": 0.5467,
+      "step": 176128
+    },
+    {
+      "epoch": 2.1669876313421237,
+      "grad_norm": 16.06914710998535,
+      "learning_rate": 1.7833026192933317e-05,
+      "loss": 0.6439,
+      "step": 176384
+    },
+    {
+      "epoch": 2.1701327780183735,
+      "grad_norm": 0.31763556599617004,
+      "learning_rate": 1.7829881075237114e-05,
+      "loss": 0.5656,
+      "step": 176640
+    },
+    {
+      "epoch": 2.1732779246946228,
+      "grad_norm": 0.7603576183319092,
+      "learning_rate": 1.7826735957540913e-05,
+      "loss": 0.6431,
+      "step": 176896
+    },
+    {
+      "epoch": 2.1764230713708725,
+      "grad_norm": 5.1138224601745605,
+      "learning_rate": 1.782359083984471e-05,
+      "loss": 0.598,
+      "step": 177152
+    },
+    {
+      "epoch": 2.179568218047122,
+      "grad_norm": 1.4138920307159424,
+      "learning_rate": 1.782044572214851e-05,
+      "loss": 0.5318,
+      "step": 177408
+    },
+    {
+      "epoch": 2.1827133647233716,
+      "grad_norm": 25.030061721801758,
+      "learning_rate": 1.781730060445231e-05,
+      "loss": 0.503,
+      "step": 177664
+    },
+    {
+      "epoch": 2.185858511399621,
+      "grad_norm": 32.86785125732422,
+      "learning_rate": 1.781415548675611e-05,
+      "loss": 0.5379,
+      "step": 177920
+    },
+    {
+      "epoch": 2.1890036580758707,
+      "grad_norm": 10.319305419921875,
+      "learning_rate": 1.7811010369059906e-05,
+      "loss": 0.6281,
+      "step": 178176
+    },
+    {
+      "epoch": 2.19214880475212,
+      "grad_norm": 34.57067108154297,
+      "learning_rate": 1.7807865251363705e-05,
+      "loss": 0.586,
+      "step": 178432
+    },
+    {
+      "epoch": 2.1952939514283694,
+      "grad_norm": 50.976505279541016,
+      "learning_rate": 1.7804720133667502e-05,
+      "loss": 0.5685,
+      "step": 178688
+    },
+    {
+      "epoch": 2.198439098104619,
+      "grad_norm": 55.41228103637695,
+      "learning_rate": 1.78015750159713e-05,
+      "loss": 0.5806,
+      "step": 178944
+    },
+    {
+      "epoch": 2.201584244780869,
+      "grad_norm": 5.2850775718688965,
+      "learning_rate": 1.77984298982751e-05,
+      "loss": 0.5684,
+      "step": 179200
+    },
+    {
+      "epoch": 2.204729391457118,
+      "grad_norm": 0.21184729039669037,
+      "learning_rate": 1.7795284780578898e-05,
+      "loss": 0.5572,
+      "step": 179456
+    },
+    {
+      "epoch": 2.2078745381333675,
+      "grad_norm": 0.21365340054035187,
+      "learning_rate": 1.7792139662882698e-05,
+      "loss": 0.5992,
+      "step": 179712
+    },
+    {
+      "epoch": 2.2110196848096173,
+      "grad_norm": 62.57757568359375,
+      "learning_rate": 1.7788994545186497e-05,
+      "loss": 0.581,
+      "step": 179968
+    },
+    {
+      "epoch": 2.2141648314858666,
+      "grad_norm": 13.238968849182129,
+      "learning_rate": 1.7785849427490297e-05,
+      "loss": 0.5375,
+      "step": 180224
+    },
+    {
+      "epoch": 2.2173099781621164,
+      "grad_norm": 2.853482961654663,
+      "learning_rate": 1.7782704309794094e-05,
+      "loss": 0.5487,
+      "step": 180480
+    },
+    {
+      "epoch": 2.2204551248383657,
+      "grad_norm": 34.67491912841797,
+      "learning_rate": 1.7779559192097893e-05,
+      "loss": 0.6092,
+      "step": 180736
+    },
+    {
+      "epoch": 2.2236002715146155,
+      "grad_norm": 0.3736776113510132,
+      "learning_rate": 1.777641407440169e-05,
+      "loss": 0.551,
+      "step": 180992
+    },
+    {
+      "epoch": 2.226745418190865,
+      "grad_norm": 0.37587207555770874,
+      "learning_rate": 1.7773268956705493e-05,
+      "loss": 0.483,
+      "step": 181248
+    },
+    {
+      "epoch": 2.2298905648671146,
+      "grad_norm": 25.50969696044922,
+      "learning_rate": 1.777012383900929e-05,
+      "loss": 0.6624,
+      "step": 181504
+    },
+    {
+      "epoch": 2.233035711543364,
+      "grad_norm": 2.3961682319641113,
+      "learning_rate": 1.776697872131309e-05,
+      "loss": 0.6188,
+      "step": 181760
+    },
+    {
+      "epoch": 2.2361808582196137,
+      "grad_norm": 20.111602783203125,
+      "learning_rate": 1.7763833603616886e-05,
+      "loss": 0.5118,
+      "step": 182016
+    },
+    {
+      "epoch": 2.239326004895863,
+      "grad_norm": 34.49458694458008,
+      "learning_rate": 1.7760688485920686e-05,
+      "loss": 0.6654,
+      "step": 182272
+    },
+    {
+      "epoch": 2.2424711515721127,
+      "grad_norm": 0.14177145063877106,
+      "learning_rate": 1.7757543368224482e-05,
+      "loss": 0.5728,
+      "step": 182528
+    },
+    {
+      "epoch": 2.245616298248362,
+      "grad_norm": 53.09233474731445,
+      "learning_rate": 1.7754398250528282e-05,
+      "loss": 0.6247,
+      "step": 182784
+    },
+    {
+      "epoch": 2.248761444924612,
+      "grad_norm": 9.360284805297852,
+      "learning_rate": 1.775125313283208e-05,
+      "loss": 0.5854,
+      "step": 183040
+    },
+    {
+      "epoch": 2.251906591600861,
+      "grad_norm": 302.94085693359375,
+      "learning_rate": 1.774810801513588e-05,
+      "loss": 0.5844,
+      "step": 183296
+    },
+    {
+      "epoch": 2.255051738277111,
+      "grad_norm": 27.76773452758789,
+      "learning_rate": 1.7744962897439678e-05,
+      "loss": 0.553,
+      "step": 183552
+    },
+    {
+      "epoch": 2.2581968849533602,
+      "grad_norm": 28.039737701416016,
+      "learning_rate": 1.7741817779743478e-05,
+      "loss": 0.6057,
+      "step": 183808
+    },
+    {
+      "epoch": 2.26134203162961,
+      "grad_norm": 26.586185455322266,
+      "learning_rate": 1.7738672662047277e-05,
+      "loss": 0.5182,
+      "step": 184064
+    },
+    {
+      "epoch": 2.2644871783058593,
+      "grad_norm": 1.7055407762527466,
+      "learning_rate": 1.7735527544351074e-05,
+      "loss": 0.5998,
+      "step": 184320
+    },
+    {
+      "epoch": 2.267632324982109,
+      "grad_norm": 31.97067642211914,
+      "learning_rate": 1.7732382426654874e-05,
+      "loss": 0.6007,
+      "step": 184576
+    },
+    {
+      "epoch": 2.2707774716583584,
+      "grad_norm": 54.70928192138672,
+      "learning_rate": 1.7729237308958673e-05,
+      "loss": 0.593,
+      "step": 184832
+    },
+    {
+      "epoch": 2.273922618334608,
+      "grad_norm": 209.2177276611328,
+      "learning_rate": 1.7726092191262473e-05,
+      "loss": 0.5217,
+      "step": 185088
+    },
+    {
+      "epoch": 2.2770677650108575,
+      "grad_norm": 9.360754013061523,
+      "learning_rate": 1.772294707356627e-05,
+      "loss": 0.5291,
+      "step": 185344
+    },
+    {
+      "epoch": 2.280212911687107,
+      "grad_norm": 5.770539283752441,
+      "learning_rate": 1.771980195587007e-05,
+      "loss": 0.5612,
+      "step": 185600
+    },
+    {
+      "epoch": 2.2833580583633566,
+      "grad_norm": 16.123714447021484,
+      "learning_rate": 1.7716656838173866e-05,
+      "loss": 0.6194,
+      "step": 185856
+    },
+    {
+      "epoch": 2.2865032050396064,
+      "grad_norm": 96.44982147216797,
+      "learning_rate": 1.7713511720477666e-05,
+      "loss": 0.6224,
+      "step": 186112
+    },
+    {
+      "epoch": 2.2896483517158557,
+      "grad_norm": 17.79292869567871,
+      "learning_rate": 1.7710366602781465e-05,
+      "loss": 0.6006,
+      "step": 186368
+    },
+    {
+      "epoch": 2.292793498392105,
+      "grad_norm": 13.114847183227539,
+      "learning_rate": 1.7707221485085265e-05,
+      "loss": 0.509,
+      "step": 186624
+    },
+    {
+      "epoch": 2.2959386450683548,
+      "grad_norm": 0.587791919708252,
+      "learning_rate": 1.770407636738906e-05,
+      "loss": 0.5518,
+      "step": 186880
+    },
+    {
+      "epoch": 2.299083791744604,
+      "grad_norm": 26.491085052490234,
+      "learning_rate": 1.770093124969286e-05,
+      "loss": 0.499,
+      "step": 187136
+    },
+    {
+      "epoch": 2.302228938420854,
+      "grad_norm": 10.07185173034668,
+      "learning_rate": 1.7697786131996658e-05,
+      "loss": 0.5946,
+      "step": 187392
+    },
+    {
+      "epoch": 2.305374085097103,
+      "grad_norm": 30.477537155151367,
+      "learning_rate": 1.7694641014300458e-05,
+      "loss": 0.5465,
+      "step": 187648
+    },
+    {
+      "epoch": 2.308519231773353,
+      "grad_norm": 10.456851959228516,
+      "learning_rate": 1.7691495896604258e-05,
+      "loss": 0.6194,
+      "step": 187904
+    },
+    {
+      "epoch": 2.3116643784496023,
+      "grad_norm": 91.97716522216797,
+      "learning_rate": 1.7688350778908057e-05,
+      "loss": 0.7026,
+      "step": 188160
+    },
+    {
+      "epoch": 2.314809525125852,
+      "grad_norm": 55.115169525146484,
+      "learning_rate": 1.7685205661211854e-05,
+      "loss": 0.54,
+      "step": 188416
+    },
+    {
+      "epoch": 2.3179546718021014,
+      "grad_norm": 1.1097732782363892,
+      "learning_rate": 1.7682060543515654e-05,
+      "loss": 0.5442,
+      "step": 188672
+    },
+    {
+      "epoch": 2.321099818478351,
+      "grad_norm": 10.730279922485352,
+      "learning_rate": 1.7678915425819453e-05,
+      "loss": 0.6763,
+      "step": 188928
+    },
+    {
+      "epoch": 2.3242449651546004,
+      "grad_norm": 52.275089263916016,
+      "learning_rate": 1.767577030812325e-05,
+      "loss": 0.5351,
+      "step": 189184
+    },
+    {
+      "epoch": 2.32739011183085,
+      "grad_norm": 25.814199447631836,
+      "learning_rate": 1.767262519042705e-05,
+      "loss": 0.5854,
+      "step": 189440
+    },
+    {
+      "epoch": 2.3305352585070995,
+      "grad_norm": 2.577927350997925,
+      "learning_rate": 1.766948007273085e-05,
+      "loss": 0.665,
+      "step": 189696
+    },
+    {
+      "epoch": 2.3336804051833493,
+      "grad_norm": 79.96446228027344,
+      "learning_rate": 1.7666334955034646e-05,
+      "loss": 0.539,
+      "step": 189952
+    },
+    {
+      "epoch": 2.3368255518595986,
+      "grad_norm": 69.51277160644531,
+      "learning_rate": 1.7663189837338446e-05,
+      "loss": 0.6262,
+      "step": 190208
+    },
+    {
+      "epoch": 2.3399706985358484,
+      "grad_norm": 2.5235626697540283,
+      "learning_rate": 1.7660044719642245e-05,
+      "loss": 0.6906,
+      "step": 190464
+    },
+    {
+      "epoch": 2.3431158452120977,
+      "grad_norm": 0.376371294260025,
+      "learning_rate": 1.7656899601946042e-05,
+      "loss": 0.4951,
+      "step": 190720
+    },
+    {
+      "epoch": 2.3462609918883475,
+      "grad_norm": 1.4877338409423828,
+      "learning_rate": 1.765375448424984e-05,
+      "loss": 0.5062,
+      "step": 190976
+    },
+    {
+      "epoch": 2.349406138564597,
+      "grad_norm": 0.3933185636997223,
+      "learning_rate": 1.765060936655364e-05,
+      "loss": 0.6049,
+      "step": 191232
+    },
+    {
+      "epoch": 2.3525512852408466,
+      "grad_norm": 12.670846939086914,
+      "learning_rate": 1.764746424885744e-05,
+      "loss": 0.5462,
+      "step": 191488
+    },
+    {
+      "epoch": 2.355696431917096,
+      "grad_norm": 53.79432678222656,
+      "learning_rate": 1.7644319131161238e-05,
+      "loss": 0.7139,
+      "step": 191744
+    },
+    {
+      "epoch": 2.3588415785933456,
+      "grad_norm": 1.4756853580474854,
+      "learning_rate": 1.7641174013465037e-05,
+      "loss": 0.5469,
+      "step": 192000
+    },
+    {
+      "epoch": 2.361986725269595,
+      "grad_norm": 83.03379821777344,
+      "learning_rate": 1.7638028895768834e-05,
+      "loss": 0.674,
+      "step": 192256
+    },
+    {
+      "epoch": 2.3651318719458443,
+      "grad_norm": 10.190073013305664,
+      "learning_rate": 1.7634883778072634e-05,
+      "loss": 0.4848,
+      "step": 192512
+    },
+    {
+      "epoch": 2.368277018622094,
+      "grad_norm": 36.900238037109375,
+      "learning_rate": 1.7631738660376433e-05,
+      "loss": 0.5795,
+      "step": 192768
+    },
+    {
+      "epoch": 2.371422165298344,
+      "grad_norm": 68.24728393554688,
+      "learning_rate": 1.7628593542680233e-05,
+      "loss": 0.5965,
+      "step": 193024
+    },
+    {
+      "epoch": 2.374567311974593,
+      "grad_norm": 79.47762298583984,
+      "learning_rate": 1.762544842498403e-05,
+      "loss": 0.5524,
+      "step": 193280
+    },
+    {
+      "epoch": 2.3777124586508425,
+      "grad_norm": 135.55921936035156,
+      "learning_rate": 1.762230330728783e-05,
+      "loss": 0.578,
+      "step": 193536
+    },
+    {
+      "epoch": 2.3808576053270922,
+      "grad_norm": 0.1675976663827896,
+      "learning_rate": 1.7619158189591626e-05,
+      "loss": 0.5636,
+      "step": 193792
+    },
+    {
+      "epoch": 2.3840027520033416,
+      "grad_norm": 18.05344581604004,
+      "learning_rate": 1.7616013071895426e-05,
+      "loss": 0.5517,
+      "step": 194048
+    },
+    {
+      "epoch": 2.3871478986795913,
+      "grad_norm": 79.73331451416016,
+      "learning_rate": 1.7612867954199226e-05,
+      "loss": 0.6532,
+      "step": 194304
+    },
+    {
+      "epoch": 2.3902930453558406,
+      "grad_norm": 49.62660217285156,
+      "learning_rate": 1.7609722836503022e-05,
+      "loss": 0.4925,
+      "step": 194560
+    },
+    {
+      "epoch": 2.3934381920320904,
+      "grad_norm": 58.083885192871094,
+      "learning_rate": 1.7606577718806822e-05,
+      "loss": 0.5288,
+      "step": 194816
+    },
+    {
+      "epoch": 2.3965833387083397,
+      "grad_norm": 10.650096893310547,
+      "learning_rate": 1.760343260111062e-05,
+      "loss": 0.5335,
+      "step": 195072
+    },
+    {
+      "epoch": 2.3997284853845895,
+      "grad_norm": 0.9964665174484253,
+      "learning_rate": 1.760028748341442e-05,
+      "loss": 0.5458,
+      "step": 195328
+    },
+    {
+      "epoch": 2.402873632060839,
+      "grad_norm": 2.2392542362213135,
+      "learning_rate": 1.7597142365718218e-05,
+      "loss": 0.5718,
+      "step": 195584
+    },
+    {
+      "epoch": 2.4060187787370886,
+      "grad_norm": 53.27199172973633,
+      "learning_rate": 1.7593997248022018e-05,
+      "loss": 0.5768,
+      "step": 195840
+    },
+    {
+      "epoch": 2.409163925413338,
+      "grad_norm": 13.158584594726562,
+      "learning_rate": 1.7590852130325814e-05,
+      "loss": 0.5442,
+      "step": 196096
+    },
+    {
+      "epoch": 2.4123090720895877,
+      "grad_norm": 56.950069427490234,
+      "learning_rate": 1.7587707012629617e-05,
+      "loss": 0.557,
+      "step": 196352
+    },
+    {
+      "epoch": 2.415454218765837,
+      "grad_norm": 4.544313430786133,
+      "learning_rate": 1.7584561894933414e-05,
+      "loss": 0.6388,
+      "step": 196608
+    },
+    {
+      "epoch": 2.4185993654420868,
+      "grad_norm": 2.138153076171875,
+      "learning_rate": 1.7581416777237213e-05,
+      "loss": 0.6396,
+      "step": 196864
+    },
+    {
+      "epoch": 2.421744512118336,
+      "grad_norm": 183.38082885742188,
+      "learning_rate": 1.757827165954101e-05,
+      "loss": 0.6649,
+      "step": 197120
+    },
+    {
+      "epoch": 2.424889658794586,
+      "grad_norm": 0.2938964366912842,
+      "learning_rate": 1.757512654184481e-05,
+      "loss": 0.6024,
+      "step": 197376
+    },
+    {
+      "epoch": 2.428034805470835,
+      "grad_norm": 32.94732666015625,
+      "learning_rate": 1.7571981424148606e-05,
+      "loss": 0.6106,
+      "step": 197632
+    },
+    {
+      "epoch": 2.431179952147085,
+      "grad_norm": 0.579232931137085,
+      "learning_rate": 1.7568836306452406e-05,
+      "loss": 0.6016,
+      "step": 197888
+    },
+    {
+      "epoch": 2.4343250988233343,
+      "grad_norm": 5.028018474578857,
+      "learning_rate": 1.7565691188756206e-05,
+      "loss": 0.4815,
+      "step": 198144
+    },
+    {
+      "epoch": 2.4374702454995836,
+      "grad_norm": 0.5712786316871643,
+      "learning_rate": 1.7562546071060005e-05,
+      "loss": 0.5892,
+      "step": 198400
+    },
+    {
+      "epoch": 2.4406153921758333,
+      "grad_norm": 2.516390085220337,
+      "learning_rate": 1.7559400953363802e-05,
+      "loss": 0.5869,
+      "step": 198656
+    },
+    {
+      "epoch": 2.443760538852083,
+      "grad_norm": 6.842493534088135,
+      "learning_rate": 1.75562558356676e-05,
+      "loss": 0.5179,
+      "step": 198912
+    },
+    {
+      "epoch": 2.4469056855283324,
+      "grad_norm": 4.345905780792236,
+      "learning_rate": 1.75531107179714e-05,
+      "loss": 0.5503,
+      "step": 199168
+    },
+    {
+      "epoch": 2.4500508322045818,
+      "grad_norm": 9.76185417175293,
+      "learning_rate": 1.7549965600275198e-05,
+      "loss": 0.6533,
+      "step": 199424
+    },
+    {
+      "epoch": 2.4531959788808315,
+      "grad_norm": 686.0620727539062,
+      "learning_rate": 1.7546820482578998e-05,
+      "loss": 0.5924,
+      "step": 199680
+    },
+    {
+      "epoch": 2.456341125557081,
+      "grad_norm": 0.8740330934524536,
+      "learning_rate": 1.7543675364882797e-05,
+      "loss": 0.5654,
+      "step": 199936
+    },
+    {
+      "epoch": 2.4594862722333306,
+      "grad_norm": 24.597936630249023,
+      "learning_rate": 1.7540530247186597e-05,
+      "loss": 0.5796,
+      "step": 200192
+    },
+    {
+      "epoch": 2.46263141890958,
+      "grad_norm": 4.8883056640625,
+      "learning_rate": 1.7537385129490394e-05,
+      "loss": 0.461,
+      "step": 200448
+    },
+    {
+      "epoch": 2.4657765655858297,
+      "grad_norm": 39.034420013427734,
+      "learning_rate": 1.7534240011794194e-05,
+      "loss": 0.5117,
+      "step": 200704
+    },
+    {
+      "epoch": 2.468921712262079,
+      "grad_norm": 56.68083190917969,
+      "learning_rate": 1.753109489409799e-05,
+      "loss": 0.5503,
+      "step": 200960
+    },
+    {
+      "epoch": 2.472066858938329,
+      "grad_norm": 4.2078704833984375,
+      "learning_rate": 1.752794977640179e-05,
+      "loss": 0.6026,
+      "step": 201216
+    },
+    {
+      "epoch": 2.475212005614578,
+      "grad_norm": 18.32750701904297,
+      "learning_rate": 1.752480465870559e-05,
+      "loss": 0.5886,
+      "step": 201472
+    },
+    {
+      "epoch": 2.478357152290828,
+      "grad_norm": 0.15369176864624023,
+      "learning_rate": 1.752165954100939e-05,
+      "loss": 0.5212,
+      "step": 201728
+    },
+    {
+      "epoch": 2.481502298967077,
+      "grad_norm": 39.12358856201172,
+      "learning_rate": 1.7518514423313186e-05,
+      "loss": 0.6062,
+      "step": 201984
+    },
+    {
+      "epoch": 2.484647445643327,
+      "grad_norm": 1.4621517658233643,
+      "learning_rate": 1.7515369305616986e-05,
+      "loss": 0.5698,
+      "step": 202240
+    },
+    {
+      "epoch": 2.4877925923195763,
+      "grad_norm": 0.6439024209976196,
+      "learning_rate": 1.7512224187920782e-05,
+      "loss": 0.6058,
+      "step": 202496
+    },
+    {
+      "epoch": 2.490937738995826,
+      "grad_norm": 135.52896118164062,
+      "learning_rate": 1.7509079070224582e-05,
+      "loss": 0.5911,
+      "step": 202752
+    },
+    {
+      "epoch": 2.4940828856720754,
+      "grad_norm": 4.252516269683838,
+      "learning_rate": 1.750593395252838e-05,
+      "loss": 0.6971,
+      "step": 203008
+    },
+    {
+      "epoch": 2.497228032348325,
+      "grad_norm": 1.9235602617263794,
+      "learning_rate": 1.750278883483218e-05,
+      "loss": 0.4947,
+      "step": 203264
+    },
+    {
+      "epoch": 2.5003731790245745,
+      "grad_norm": 40.87093734741211,
+      "learning_rate": 1.7499643717135978e-05,
+      "loss": 0.6329,
+      "step": 203520
+    },
+    {
+      "epoch": 2.503518325700824,
+      "grad_norm": 0.918137788772583,
+      "learning_rate": 1.7496498599439778e-05,
+      "loss": 0.5642,
+      "step": 203776
+    },
+    {
+      "epoch": 2.5066634723770735,
+      "grad_norm": 1.5068498849868774,
+      "learning_rate": 1.7493353481743577e-05,
+      "loss": 0.5068,
+      "step": 204032
+    },
+    {
+      "epoch": 2.509808619053323,
+      "grad_norm": 19.423437118530273,
+      "learning_rate": 1.7490208364047374e-05,
+      "loss": 0.5774,
+      "step": 204288
+    },
+    {
+      "epoch": 2.5129537657295726,
+      "grad_norm": 20.438018798828125,
+      "learning_rate": 1.7487063246351174e-05,
+      "loss": 0.5597,
+      "step": 204544
+    },
+    {
+      "epoch": 2.5160989124058224,
+      "grad_norm": 510.6248779296875,
+      "learning_rate": 1.7483918128654973e-05,
+      "loss": 0.6073,
+      "step": 204800
+    },
+    {
+      "epoch": 2.5192440590820717,
+      "grad_norm": 1.4464325904846191,
+      "learning_rate": 1.748077301095877e-05,
+      "loss": 0.6472,
+      "step": 205056
+    },
+    {
+      "epoch": 2.522389205758321,
+      "grad_norm": 93.39960479736328,
+      "learning_rate": 1.747762789326257e-05,
+      "loss": 0.6295,
+      "step": 205312
+    },
+    {
+      "epoch": 2.525534352434571,
+      "grad_norm": 28.367328643798828,
+      "learning_rate": 1.747448277556637e-05,
+      "loss": 0.6148,
+      "step": 205568
+    },
+    {
+      "epoch": 2.5286794991108206,
+      "grad_norm": 20.399938583374023,
+      "learning_rate": 1.7471337657870166e-05,
+      "loss": 0.5694,
+      "step": 205824
+    },
+    {
+      "epoch": 2.53182464578707,
+      "grad_norm": 9.061676025390625,
+      "learning_rate": 1.7468192540173966e-05,
+      "loss": 0.5552,
+      "step": 206080
+    },
+    {
+      "epoch": 2.534969792463319,
+      "grad_norm": 3.76851487159729,
+      "learning_rate": 1.7465047422477765e-05,
+      "loss": 0.5089,
+      "step": 206336
+    },
+    {
+      "epoch": 2.538114939139569,
+      "grad_norm": 14.37346363067627,
+      "learning_rate": 1.7461902304781565e-05,
+      "loss": 0.5758,
+      "step": 206592
+    },
+    {
+      "epoch": 2.5412600858158187,
+      "grad_norm": 228.0797576904297,
+      "learning_rate": 1.7458757187085362e-05,
+      "loss": 0.5911,
+      "step": 206848
+    },
+    {
+      "epoch": 2.544405232492068,
+      "grad_norm": 72.00253295898438,
+      "learning_rate": 1.745561206938916e-05,
+      "loss": 0.5609,
+      "step": 207104
+    },
+    {
+      "epoch": 2.5475503791683174,
+      "grad_norm": 344.53436279296875,
+      "learning_rate": 1.7452466951692958e-05,
+      "loss": 0.6061,
+      "step": 207360
+    },
+    {
+      "epoch": 2.550695525844567,
+      "grad_norm": 1.4079500436782837,
+      "learning_rate": 1.7449321833996758e-05,
+      "loss": 0.5746,
+      "step": 207616
+    },
+    {
+      "epoch": 2.5538406725208165,
+      "grad_norm": 0.46861767768859863,
+      "learning_rate": 1.7446176716300554e-05,
+      "loss": 0.5837,
+      "step": 207872
+    },
+    {
+      "epoch": 2.5569858191970662,
+      "grad_norm": 7.074849605560303,
+      "learning_rate": 1.7443031598604357e-05,
+      "loss": 0.6091,
+      "step": 208128
+    },
+    {
+      "epoch": 2.5601309658733156,
+      "grad_norm": 15.70569133758545,
+      "learning_rate": 1.7439886480908154e-05,
+      "loss": 0.5322,
+      "step": 208384
+    },
+    {
+      "epoch": 2.5632761125495653,
+      "grad_norm": 386.5798645019531,
+      "learning_rate": 1.7436741363211954e-05,
+      "loss": 0.6132,
+      "step": 208640
+    },
+    {
+      "epoch": 2.5664212592258147,
+      "grad_norm": 83.68233489990234,
+      "learning_rate": 1.743359624551575e-05,
+      "loss": 0.5068,
+      "step": 208896
+    },
+    {
+      "epoch": 2.5695664059020644,
+      "grad_norm": 15.298114776611328,
+      "learning_rate": 1.743045112781955e-05,
+      "loss": 0.5663,
+      "step": 209152
+    },
+    {
+      "epoch": 2.5727115525783137,
+      "grad_norm": 53.673301696777344,
+      "learning_rate": 1.742730601012335e-05,
+      "loss": 0.593,
+      "step": 209408
+    },
+    {
+      "epoch": 2.5758566992545635,
+      "grad_norm": 4.221065521240234,
+      "learning_rate": 1.742416089242715e-05,
+      "loss": 0.6436,
+      "step": 209664
+    },
+    {
+      "epoch": 2.579001845930813,
+      "grad_norm": 2.032090425491333,
+      "learning_rate": 1.7421015774730946e-05,
+      "loss": 0.5885,
+      "step": 209920
+    },
+    {
+      "epoch": 2.5821469926070626,
+      "grad_norm": 0.4689025580883026,
+      "learning_rate": 1.7417870657034746e-05,
+      "loss": 0.5897,
+      "step": 210176
+    },
+    {
+      "epoch": 2.585292139283312,
+      "grad_norm": 103.12809753417969,
+      "learning_rate": 1.7414725539338545e-05,
+      "loss": 0.5506,
+      "step": 210432
+    },
+    {
+      "epoch": 2.5884372859595617,
+      "grad_norm": 3.140535354614258,
+      "learning_rate": 1.7411580421642342e-05,
+      "loss": 0.6368,
+      "step": 210688
+    },
+    {
+      "epoch": 2.591582432635811,
+      "grad_norm": 23.34222412109375,
+      "learning_rate": 1.740843530394614e-05,
+      "loss": 0.535,
+      "step": 210944
+    },
+    {
+      "epoch": 2.5947275793120603,
+      "grad_norm": 2.0860018730163574,
+      "learning_rate": 1.7405290186249938e-05,
+      "loss": 0.6383,
+      "step": 211200
+    },
+    {
+      "epoch": 2.59787272598831,
+      "grad_norm": 0.22698965668678284,
+      "learning_rate": 1.740214506855374e-05,
+      "loss": 0.629,
+      "step": 211456
+    },
+    {
+      "epoch": 2.60101787266456,
+      "grad_norm": 66.57173919677734,
+      "learning_rate": 1.7398999950857538e-05,
+      "loss": 0.5836,
+      "step": 211712
+    },
+    {
+      "epoch": 2.604163019340809,
+      "grad_norm": 19.26184844970703,
+      "learning_rate": 1.7395854833161337e-05,
+      "loss": 0.5722,
+      "step": 211968
+    },
+    {
+      "epoch": 2.6073081660170585,
+      "grad_norm": 35.13220977783203,
+      "learning_rate": 1.7392709715465134e-05,
+      "loss": 0.4966,
+      "step": 212224
+    },
+    {
+      "epoch": 2.6104533126933083,
+      "grad_norm": 1.4526447057724,
+      "learning_rate": 1.7389564597768934e-05,
+      "loss": 0.5108,
+      "step": 212480
+    },
+    {
+      "epoch": 2.613598459369558,
+      "grad_norm": 0.8906441330909729,
+      "learning_rate": 1.738641948007273e-05,
+      "loss": 0.5636,
+      "step": 212736
+    },
+    {
+      "epoch": 2.6167436060458074,
+      "grad_norm": 39.04792404174805,
+      "learning_rate": 1.738327436237653e-05,
+      "loss": 0.5521,
+      "step": 212992
+    },
+    {
+      "epoch": 2.6198887527220567,
+      "grad_norm": 13.711865425109863,
+      "learning_rate": 1.738012924468033e-05,
+      "loss": 0.5684,
+      "step": 213248
+    },
+    {
+      "epoch": 2.6230338993983064,
+      "grad_norm": 76.37763214111328,
+      "learning_rate": 1.737698412698413e-05,
+      "loss": 0.6134,
+      "step": 213504
+    },
+    {
+      "epoch": 2.626179046074556,
+      "grad_norm": 81.92778015136719,
+      "learning_rate": 1.7373839009287926e-05,
+      "loss": 0.6392,
+      "step": 213760
+    },
+    {
+      "epoch": 2.6293241927508055,
+      "grad_norm": 0.5622560381889343,
+      "learning_rate": 1.7370693891591726e-05,
+      "loss": 0.5139,
+      "step": 214016
+    },
+    {
+      "epoch": 2.632469339427055,
+      "grad_norm": 92.07581329345703,
+      "learning_rate": 1.7367548773895526e-05,
+      "loss": 0.5771,
+      "step": 214272
+    },
+    {
+      "epoch": 2.6356144861033046,
+      "grad_norm": 7.411790370941162,
+      "learning_rate": 1.7364403656199322e-05,
+      "loss": 0.5158,
+      "step": 214528
+    },
+    {
+      "epoch": 2.638759632779554,
+      "grad_norm": 92.85969543457031,
+      "learning_rate": 1.7361258538503122e-05,
+      "loss": 0.6132,
+      "step": 214784
+    },
+    {
+      "epoch": 2.6419047794558037,
+      "grad_norm": 44.10087585449219,
+      "learning_rate": 1.735811342080692e-05,
+      "loss": 0.6615,
+      "step": 215040
+    },
+    {
+      "epoch": 2.645049926132053,
+      "grad_norm": 88.57605743408203,
+      "learning_rate": 1.735496830311072e-05,
+      "loss": 0.6032,
+      "step": 215296
+    },
+    {
+      "epoch": 2.648195072808303,
+      "grad_norm": 25.04104232788086,
+      "learning_rate": 1.7351823185414518e-05,
+      "loss": 0.6359,
+      "step": 215552
+    },
+    {
+      "epoch": 2.651340219484552,
+      "grad_norm": 5.337998867034912,
+      "learning_rate": 1.7348678067718318e-05,
+      "loss": 0.5871,
+      "step": 215808
+    },
+    {
+      "epoch": 2.654485366160802,
+      "grad_norm": 48.22161865234375,
+      "learning_rate": 1.7345532950022114e-05,
+      "loss": 0.6114,
+      "step": 216064
+    },
+    {
+      "epoch": 2.657630512837051,
+      "grad_norm": 1.0288362503051758,
+      "learning_rate": 1.7342387832325914e-05,
+      "loss": 0.658,
+      "step": 216320
+    },
+    {
+      "epoch": 2.660775659513301,
+      "grad_norm": 1.7818485498428345,
+      "learning_rate": 1.7339242714629714e-05,
+      "loss": 0.5906,
+      "step": 216576
+    },
+    {
+      "epoch": 2.6639208061895503,
+      "grad_norm": 148.84921264648438,
+      "learning_rate": 1.7336097596933513e-05,
+      "loss": 0.6164,
+      "step": 216832
+    },
+    {
+      "epoch": 2.6670659528658,
+      "grad_norm": 546.583251953125,
+      "learning_rate": 1.733295247923731e-05,
+      "loss": 0.6501,
+      "step": 217088
+    },
+    {
+      "epoch": 2.6702110995420494,
+      "grad_norm": 19.556991577148438,
+      "learning_rate": 1.732980736154111e-05,
+      "loss": 0.5601,
+      "step": 217344
+    },
+    {
+      "epoch": 2.673356246218299,
+      "grad_norm": 0.1624416708946228,
+      "learning_rate": 1.7326662243844906e-05,
+      "loss": 0.5389,
+      "step": 217600
+    },
+    {
+      "epoch": 2.6765013928945485,
+      "grad_norm": 0.8040470480918884,
+      "learning_rate": 1.7323517126148706e-05,
+      "loss": 0.6163,
+      "step": 217856
+    },
+    {
+      "epoch": 2.679646539570798,
+      "grad_norm": 23.409908294677734,
+      "learning_rate": 1.7320372008452506e-05,
+      "loss": 0.6085,
+      "step": 218112
+    },
+    {
+      "epoch": 2.6827916862470476,
+      "grad_norm": 36.47003173828125,
+      "learning_rate": 1.7317226890756305e-05,
+      "loss": 0.5895,
+      "step": 218368
+    },
+    {
+      "epoch": 2.6859368329232973,
+      "grad_norm": 0.584807276725769,
+      "learning_rate": 1.7314081773060102e-05,
+      "loss": 0.5938,
+      "step": 218624
+    },
+    {
+      "epoch": 2.6890819795995466,
+      "grad_norm": 77.03465270996094,
+      "learning_rate": 1.73109366553639e-05,
+      "loss": 0.5505,
+      "step": 218880
+    },
+    {
+      "epoch": 2.692227126275796,
+      "grad_norm": 13.01475715637207,
+      "learning_rate": 1.7307791537667698e-05,
+      "loss": 0.6421,
+      "step": 219136
+    },
+    {
+      "epoch": 2.6953722729520457,
+      "grad_norm": 14.55999755859375,
+      "learning_rate": 1.7304646419971498e-05,
+      "loss": 0.5396,
+      "step": 219392
+    },
+    {
+      "epoch": 2.6985174196282955,
+      "grad_norm": 58.929542541503906,
+      "learning_rate": 1.7301501302275298e-05,
+      "loss": 0.6403,
+      "step": 219648
+    },
+    {
+      "epoch": 2.701662566304545,
+      "grad_norm": 60.64753341674805,
+      "learning_rate": 1.7298356184579098e-05,
+      "loss": 0.5928,
+      "step": 219904
+    },
+    {
+      "epoch": 2.704807712980794,
+      "grad_norm": 48.01443862915039,
+      "learning_rate": 1.7295211066882894e-05,
+      "loss": 0.614,
+      "step": 220160
+    },
+    {
+      "epoch": 2.707952859657044,
+      "grad_norm": 84.26978302001953,
+      "learning_rate": 1.7292065949186694e-05,
+      "loss": 0.5366,
+      "step": 220416
+    },
+    {
+      "epoch": 2.7110980063332937,
+      "grad_norm": 1.246039867401123,
+      "learning_rate": 1.7288920831490494e-05,
+      "loss": 0.5209,
+      "step": 220672
+    },
+    {
+      "epoch": 2.714243153009543,
+      "grad_norm": 123.53519439697266,
+      "learning_rate": 1.728577571379429e-05,
+      "loss": 0.5139,
+      "step": 220928
+    },
+    {
+      "epoch": 2.7173882996857923,
+      "grad_norm": 12.696935653686523,
+      "learning_rate": 1.728263059609809e-05,
+      "loss": 0.5732,
+      "step": 221184
+    },
+    {
+      "epoch": 2.720533446362042,
+      "grad_norm": 32.884822845458984,
+      "learning_rate": 1.727948547840189e-05,
+      "loss": 0.6007,
+      "step": 221440
+    },
+    {
+      "epoch": 2.7236785930382914,
+      "grad_norm": 3.320554256439209,
+      "learning_rate": 1.727634036070569e-05,
+      "loss": 0.5718,
+      "step": 221696
+    },
+    {
+      "epoch": 2.726823739714541,
+      "grad_norm": 1.9137370586395264,
+      "learning_rate": 1.7273195243009486e-05,
+      "loss": 0.5854,
+      "step": 221952
+    },
+    {
+      "epoch": 2.7299688863907905,
+      "grad_norm": 6.873824119567871,
+      "learning_rate": 1.7270050125313286e-05,
+      "loss": 0.5018,
+      "step": 222208
+    },
+    {
+      "epoch": 2.7331140330670403,
+      "grad_norm": 53.9407958984375,
+      "learning_rate": 1.7266905007617082e-05,
+      "loss": 0.5601,
+      "step": 222464
+    },
+    {
+      "epoch": 2.7362591797432896,
+      "grad_norm": 6.90484094619751,
+      "learning_rate": 1.7263759889920882e-05,
+      "loss": 0.5471,
+      "step": 222720
+    },
+    {
+      "epoch": 2.7394043264195393,
+      "grad_norm": 70.38066864013672,
+      "learning_rate": 1.7260614772224678e-05,
+      "loss": 0.6211,
+      "step": 222976
+    },
+    {
+      "epoch": 2.7425494730957887,
+      "grad_norm": 4.555383205413818,
+      "learning_rate": 1.725746965452848e-05,
+      "loss": 0.5571,
+      "step": 223232
+    },
+    {
+      "epoch": 2.7456946197720384,
+      "grad_norm": 4.549954414367676,
+      "learning_rate": 1.7254324536832278e-05,
+      "loss": 0.5378,
+      "step": 223488
+    },
+    {
+      "epoch": 2.7488397664482878,
+      "grad_norm": 117.39866638183594,
+      "learning_rate": 1.7251179419136078e-05,
+      "loss": 0.6538,
+      "step": 223744
+    },
+    {
+      "epoch": 2.7519849131245375,
+      "grad_norm": 61.44218063354492,
+      "learning_rate": 1.7248034301439874e-05,
+      "loss": 0.6754,
+      "step": 224000
+    },
+    {
+      "epoch": 2.755130059800787,
+      "grad_norm": 5.581026554107666,
+      "learning_rate": 1.7244889183743674e-05,
+      "loss": 0.5284,
+      "step": 224256
+    },
+    {
+      "epoch": 2.7582752064770366,
+      "grad_norm": 0.45108968019485474,
+      "learning_rate": 1.7241744066047474e-05,
+      "loss": 0.5066,
+      "step": 224512
+    },
+    {
+      "epoch": 2.761420353153286,
+      "grad_norm": 45.785667419433594,
+      "learning_rate": 1.7238598948351273e-05,
+      "loss": 0.5734,
+      "step": 224768
+    },
+    {
+      "epoch": 2.7645654998295353,
+      "grad_norm": 67.40306091308594,
+      "learning_rate": 1.723545383065507e-05,
+      "loss": 0.5598,
+      "step": 225024
+    },
+    {
+      "epoch": 2.767710646505785,
+      "grad_norm": 15.22658920288086,
+      "learning_rate": 1.723230871295887e-05,
+      "loss": 0.6087,
+      "step": 225280
+    },
+    {
+      "epoch": 2.770855793182035,
+      "grad_norm": 33.362953186035156,
+      "learning_rate": 1.722916359526267e-05,
+      "loss": 0.6499,
+      "step": 225536
+    },
+    {
+      "epoch": 2.774000939858284,
+      "grad_norm": 220.8497772216797,
+      "learning_rate": 1.7226018477566466e-05,
+      "loss": 0.625,
+      "step": 225792
+    },
+    {
+      "epoch": 2.7771460865345334,
+      "grad_norm": 140.84242248535156,
+      "learning_rate": 1.7222873359870266e-05,
+      "loss": 0.5851,
+      "step": 226048
+    },
+    {
+      "epoch": 2.780291233210783,
+      "grad_norm": 0.5248227119445801,
+      "learning_rate": 1.7219728242174062e-05,
+      "loss": 0.5649,
+      "step": 226304
+    },
+    {
+      "epoch": 2.783436379887033,
+      "grad_norm": 5.1078596115112305,
+      "learning_rate": 1.7216583124477865e-05,
+      "loss": 0.5151,
+      "step": 226560
+    },
+    {
+      "epoch": 2.7865815265632823,
+      "grad_norm": 132.41030883789062,
+      "learning_rate": 1.7213438006781662e-05,
+      "loss": 0.6367,
+      "step": 226816
+    },
+    {
+      "epoch": 2.7897266732395316,
+      "grad_norm": 4.824883937835693,
+      "learning_rate": 1.721029288908546e-05,
+      "loss": 0.5625,
+      "step": 227072
+    },
+    {
+      "epoch": 2.7928718199157814,
+      "grad_norm": 31.026464462280273,
+      "learning_rate": 1.7207147771389258e-05,
+      "loss": 0.7072,
+      "step": 227328
+    },
+    {
+      "epoch": 2.796016966592031,
+      "grad_norm": 23.662321090698242,
+      "learning_rate": 1.7204002653693058e-05,
+      "loss": 0.5188,
+      "step": 227584
+    },
+    {
+      "epoch": 2.7991621132682805,
+      "grad_norm": 2.1527605056762695,
+      "learning_rate": 1.7200857535996854e-05,
+      "loss": 0.6014,
+      "step": 227840
+    },
+    {
+      "epoch": 2.80230725994453,
+      "grad_norm": 37.09303665161133,
+      "learning_rate": 1.7197712418300657e-05,
+      "loss": 0.5671,
+      "step": 228096
+    },
+    {
+      "epoch": 2.8054524066207795,
+      "grad_norm": 38.7642707824707,
+      "learning_rate": 1.7194567300604454e-05,
+      "loss": 0.5931,
+      "step": 228352
+    },
+    {
+      "epoch": 2.808597553297029,
+      "grad_norm": 32.17414855957031,
+      "learning_rate": 1.7191422182908254e-05,
+      "loss": 0.58,
+      "step": 228608
+    },
+    {
+      "epoch": 2.8117426999732786,
+      "grad_norm": 38.99076843261719,
+      "learning_rate": 1.718827706521205e-05,
+      "loss": 0.5142,
+      "step": 228864
+    },
+    {
+      "epoch": 2.814887846649528,
+      "grad_norm": 4.88992166519165,
+      "learning_rate": 1.718513194751585e-05,
+      "loss": 0.5777,
+      "step": 229120
+    },
+    {
+      "epoch": 2.8180329933257777,
+      "grad_norm": 11.647442817687988,
+      "learning_rate": 1.718198682981965e-05,
+      "loss": 0.5391,
+      "step": 229376
+    },
+    {
+      "epoch": 2.821178140002027,
+      "grad_norm": 2.532973051071167,
+      "learning_rate": 1.7178841712123446e-05,
+      "loss": 0.5705,
+      "step": 229632
+    },
+    {
+      "epoch": 2.824323286678277,
+      "grad_norm": 8.589701652526855,
+      "learning_rate": 1.7175696594427246e-05,
+      "loss": 0.5046,
+      "step": 229888
+    },
+    {
+      "epoch": 2.827468433354526,
+      "grad_norm": 1.0302560329437256,
+      "learning_rate": 1.7172551476731046e-05,
+      "loss": 0.6222,
+      "step": 230144
+    },
+    {
+      "epoch": 2.830613580030776,
+      "grad_norm": 103.66107940673828,
+      "learning_rate": 1.7169406359034842e-05,
+      "loss": 0.5188,
+      "step": 230400
+    },
+    {
+      "epoch": 2.833758726707025,
+      "grad_norm": 7.879088401794434,
+      "learning_rate": 1.7166261241338642e-05,
+      "loss": 0.5781,
+      "step": 230656
+    },
+    {
+      "epoch": 2.836903873383275,
+      "grad_norm": 38.58749008178711,
+      "learning_rate": 1.716311612364244e-05,
+      "loss": 0.5144,
+      "step": 230912
+    },
+    {
+      "epoch": 2.8400490200595243,
+      "grad_norm": 2.055776357650757,
+      "learning_rate": 1.7159971005946238e-05,
+      "loss": 0.5865,
+      "step": 231168
+    },
+    {
+      "epoch": 2.843194166735774,
+      "grad_norm": 15.951428413391113,
+      "learning_rate": 1.7156825888250038e-05,
+      "loss": 0.5366,
+      "step": 231424
+    },
+    {
+      "epoch": 2.8463393134120234,
+      "grad_norm": 7.423271656036377,
+      "learning_rate": 1.7153680770553838e-05,
+      "loss": 0.6002,
+      "step": 231680
+    },
+    {
+      "epoch": 2.8494844600882727,
+      "grad_norm": 72.87504577636719,
+      "learning_rate": 1.7150535652857637e-05,
+      "loss": 0.5843,
+      "step": 231936
+    },
+    {
+      "epoch": 2.8526296067645225,
+      "grad_norm": 2.404435873031616,
+      "learning_rate": 1.7147390535161434e-05,
+      "loss": 0.5868,
+      "step": 232192
+    },
+    {
+      "epoch": 2.8557747534407723,
+      "grad_norm": 41.06608200073242,
+      "learning_rate": 1.7144245417465234e-05,
+      "loss": 0.6576,
+      "step": 232448
+    },
+    {
+      "epoch": 2.8589199001170216,
+      "grad_norm": 48.31210708618164,
+      "learning_rate": 1.714110029976903e-05,
+      "loss": 0.6732,
+      "step": 232704
+    },
+    {
+      "epoch": 2.862065046793271,
+      "grad_norm": 20.90900993347168,
+      "learning_rate": 1.713795518207283e-05,
+      "loss": 0.6019,
+      "step": 232960
+    },
+    {
+      "epoch": 2.8652101934695207,
+      "grad_norm": 5.697980880737305,
+      "learning_rate": 1.713481006437663e-05,
+      "loss": 0.5865,
+      "step": 233216
+    },
+    {
+      "epoch": 2.8683553401457704,
+      "grad_norm": 187.47190856933594,
+      "learning_rate": 1.713166494668043e-05,
+      "loss": 0.5561,
+      "step": 233472
+    },
+    {
+      "epoch": 2.8715004868220197,
+      "grad_norm": 18.154407501220703,
+      "learning_rate": 1.7128519828984226e-05,
+      "loss": 0.6007,
+      "step": 233728
+    },
+    {
+      "epoch": 2.874645633498269,
+      "grad_norm": 12.933354377746582,
+      "learning_rate": 1.7125374711288026e-05,
+      "loss": 0.5769,
+      "step": 233984
+    },
+    {
+      "epoch": 2.877790780174519,
+      "grad_norm": 28.97356414794922,
+      "learning_rate": 1.7122229593591822e-05,
+      "loss": 0.6301,
+      "step": 234240
+    },
+    {
+      "epoch": 2.880935926850768,
+      "grad_norm": 113.57801818847656,
+      "learning_rate": 1.7119084475895622e-05,
+      "loss": 0.5929,
+      "step": 234496
+    },
+    {
+      "epoch": 2.884081073527018,
+      "grad_norm": 1.659719705581665,
+      "learning_rate": 1.7115939358199422e-05,
+      "loss": 0.4997,
+      "step": 234752
+    },
+    {
+      "epoch": 2.8872262202032672,
+      "grad_norm": 9.596588134765625,
+      "learning_rate": 1.711279424050322e-05,
+      "loss": 0.5403,
+      "step": 235008
+    },
+    {
+      "epoch": 2.890371366879517,
+      "grad_norm": 2.0734522342681885,
+      "learning_rate": 1.7109649122807018e-05,
+      "loss": 0.6429,
+      "step": 235264
+    },
+    {
+      "epoch": 2.8935165135557663,
+      "grad_norm": 65.2406997680664,
+      "learning_rate": 1.7106504005110818e-05,
+      "loss": 0.5299,
+      "step": 235520
+    },
+    {
+      "epoch": 2.896661660232016,
+      "grad_norm": 87.21231079101562,
+      "learning_rate": 1.7103358887414618e-05,
+      "loss": 0.4749,
+      "step": 235776
+    },
+    {
+      "epoch": 2.8998068069082654,
+      "grad_norm": 406.5826721191406,
+      "learning_rate": 1.7100213769718414e-05,
+      "loss": 0.5804,
+      "step": 236032
+    },
+    {
+      "epoch": 2.902951953584515,
+      "grad_norm": 0.7037609219551086,
+      "learning_rate": 1.7097068652022214e-05,
+      "loss": 0.6237,
+      "step": 236288
+    },
+    {
+      "epoch": 2.9060971002607645,
+      "grad_norm": 1.434501051902771,
+      "learning_rate": 1.7093923534326014e-05,
+      "loss": 0.5232,
+      "step": 236544
+    },
+    {
+      "epoch": 2.9092422469370143,
+      "grad_norm": 29.717830657958984,
+      "learning_rate": 1.7090778416629813e-05,
+      "loss": 0.527,
+      "step": 236800
+    },
+    {
+      "epoch": 2.9123873936132636,
+      "grad_norm": 178.4339141845703,
+      "learning_rate": 1.708763329893361e-05,
+      "loss": 0.5354,
+      "step": 237056
+    },
+    {
+      "epoch": 2.9155325402895134,
+      "grad_norm": 3.884004592895508,
+      "learning_rate": 1.708448818123741e-05,
+      "loss": 0.636,
+      "step": 237312
+    },
+    {
+      "epoch": 2.9186776869657627,
+      "grad_norm": 63.62284851074219,
+      "learning_rate": 1.7081343063541206e-05,
+      "loss": 0.5072,
+      "step": 237568
+    },
+    {
+      "epoch": 2.921822833642012,
+      "grad_norm": 43.29448699951172,
+      "learning_rate": 1.7078197945845006e-05,
+      "loss": 0.6166,
+      "step": 237824
+    },
+    {
+      "epoch": 2.9249679803182618,
+      "grad_norm": 72.13641357421875,
+      "learning_rate": 1.7075052828148802e-05,
+      "loss": 0.5611,
+      "step": 238080
+    },
+    {
+      "epoch": 2.9281131269945115,
+      "grad_norm": 117.06077575683594,
+      "learning_rate": 1.7071907710452605e-05,
+      "loss": 0.5221,
+      "step": 238336
+    },
+    {
+      "epoch": 2.931258273670761,
+      "grad_norm": 67.11541748046875,
+      "learning_rate": 1.7068762592756402e-05,
+      "loss": 0.5006,
+      "step": 238592
+    },
+    {
+      "epoch": 2.93440342034701,
+      "grad_norm": 0.4315754473209381,
+      "learning_rate": 1.7065617475060202e-05,
+      "loss": 0.5729,
+      "step": 238848
+    },
+    {
+      "epoch": 2.93754856702326,
+      "grad_norm": 0.7335332036018372,
+      "learning_rate": 1.7062472357363998e-05,
+      "loss": 0.5596,
+      "step": 239104
+    },
+    {
+      "epoch": 2.9406937136995097,
+      "grad_norm": 59.091670989990234,
+      "learning_rate": 1.7059327239667798e-05,
+      "loss": 0.5833,
+      "step": 239360
+    },
+    {
+      "epoch": 2.943838860375759,
+      "grad_norm": 26.088376998901367,
+      "learning_rate": 1.7056182121971598e-05,
+      "loss": 0.6299,
+      "step": 239616
+    },
+    {
+      "epoch": 2.9469840070520084,
+      "grad_norm": 0.3894733190536499,
+      "learning_rate": 1.7053037004275398e-05,
+      "loss": 0.5532,
+      "step": 239872
+    },
+    {
+      "epoch": 2.950129153728258,
+      "grad_norm": 0.1327989399433136,
+      "learning_rate": 1.7049891886579194e-05,
+      "loss": 0.5185,
+      "step": 240128
+    },
+    {
+      "epoch": 2.953274300404508,
+      "grad_norm": 66.7145004272461,
+      "learning_rate": 1.7046746768882994e-05,
+      "loss": 0.5133,
+      "step": 240384
+    },
+    {
+      "epoch": 2.956419447080757,
+      "grad_norm": 288.0764465332031,
+      "learning_rate": 1.7043601651186794e-05,
+      "loss": 0.5056,
+      "step": 240640
+    },
+    {
+      "epoch": 2.9595645937570065,
+      "grad_norm": 1.068016529083252,
+      "learning_rate": 1.704045653349059e-05,
+      "loss": 0.5286,
+      "step": 240896
+    },
+    {
+      "epoch": 2.9627097404332563,
+      "grad_norm": 0.3362673223018646,
+      "learning_rate": 1.703731141579439e-05,
+      "loss": 0.592,
+      "step": 241152
+    },
+    {
+      "epoch": 2.9658548871095056,
+      "grad_norm": 10.113340377807617,
+      "learning_rate": 1.7034166298098186e-05,
+      "loss": 0.5605,
+      "step": 241408
+    },
+    {
+      "epoch": 2.9690000337857554,
+      "grad_norm": 282.2839050292969,
+      "learning_rate": 1.7031021180401986e-05,
+      "loss": 0.5699,
+      "step": 241664
+    },
+    {
+      "epoch": 2.9721451804620047,
+      "grad_norm": 3.803244113922119,
+      "learning_rate": 1.7027876062705786e-05,
+      "loss": 0.6502,
+      "step": 241920
+    },
+    {
+      "epoch": 2.9752903271382545,
+      "grad_norm": 2.9723453521728516,
+      "learning_rate": 1.7024730945009586e-05,
+      "loss": 0.5807,
+      "step": 242176
+    },
+    {
+      "epoch": 2.978435473814504,
+      "grad_norm": 56.946449279785156,
+      "learning_rate": 1.7021585827313382e-05,
+      "loss": 0.6497,
+      "step": 242432
+    },
+    {
+      "epoch": 2.9815806204907536,
+      "grad_norm": 9.507156372070312,
+      "learning_rate": 1.7018440709617182e-05,
+      "loss": 0.5437,
+      "step": 242688
+    },
+    {
+      "epoch": 2.984725767167003,
+      "grad_norm": 54.93104553222656,
+      "learning_rate": 1.7015295591920978e-05,
+      "loss": 0.5746,
+      "step": 242944
+    },
+    {
+      "epoch": 2.9878709138432527,
+      "grad_norm": 608.7354125976562,
+      "learning_rate": 1.701215047422478e-05,
+      "loss": 0.5767,
+      "step": 243200
+    },
+    {
+      "epoch": 2.991016060519502,
+      "grad_norm": 0.44110745191574097,
+      "learning_rate": 1.7009005356528578e-05,
+      "loss": 0.4886,
+      "step": 243456
+    },
+    {
+      "epoch": 2.9941612071957517,
+      "grad_norm": 5.9228010177612305,
+      "learning_rate": 1.7005860238832378e-05,
+      "loss": 0.7261,
+      "step": 243712
+    },
+    {
+      "epoch": 2.997306353872001,
+      "grad_norm": 10.245580673217773,
+      "learning_rate": 1.7002715121136174e-05,
+      "loss": 0.5409,
+      "step": 243968
+    },
+    {
+      "epoch": 3.0,
+      "eval_AGE": {
+        "f1": 0.9627531597380844,
+        "number": 32192,
+        "precision": 0.9442619033395065,
+        "recall": 0.9819831013916501
+      },
+      "eval_BUILDINGNUM": {
+        "f1": 0.9776639978007629,
+        "number": 28889,
+        "precision": 0.9705932521406885,
+        "recall": 0.9848385198518467
+      },
+      "eval_CITY": {
+        "f1": 0.9662582155718513,
+        "number": 41669,
+        "precision": 0.9571462208617848,
+        "recall": 0.9755453694593103
+      },
+      "eval_DATE": {
+        "f1": 0.9970651342886034,
+        "number": 80940,
+        "precision": 0.9951752658526979,
+        "recall": 0.9989621942179392
+      },
+      "eval_EMAIL": {
+        "f1": 0.9963071660414764,
+        "number": 51445,
+        "precision": 0.9961909944225276,
+        "recall": 0.9964233647584799
+      },
+      "eval_GENDER": {
+        "f1": 0.9727332930235236,
+        "number": 17457,
+        "precision": 0.9778858399907375,
+        "recall": 0.9676347596952511
+      },
+      "eval_GIVENNAME": {
+        "f1": 0.9249527325780471,
+        "number": 100350,
+        "precision": 0.9212360250289138,
+        "recall": 0.9286995515695067
+      },
+      "eval_IDCARDNUM": {
+        "f1": 0.9027978721036348,
+        "number": 22068,
+        "precision": 0.8871779167942605,
+        "recall": 0.9189777052746058
+      },
+      "eval_PASSPORTNUM": {
+        "f1": 0.9174223279322135,
+        "number": 12959,
+        "precision": 0.8848662986594021,
+        "recall": 0.9524654680145073
+      },
+      "eval_SEX": {
+        "f1": 0.9671593330402137,
+        "number": 14570,
+        "precision": 0.9533906781356272,
+        "recall": 0.9813315030885381
+      },
+      "eval_SOCIALNUM": {
+        "f1": 0.9133946670824886,
+        "number": 15327,
+        "precision": 0.8748954474847652,
+        "recall": 0.9554381157434593
+      },
+      "eval_STREET": {
+        "f1": 0.9728207380040277,
+        "number": 32803,
+        "precision": 0.9664259927797834,
+        "recall": 0.9793006737188672
+      },
+      "eval_SURNAME": {
+        "f1": 0.9181983461380564,
+        "number": 85192,
+        "precision": 0.9175203356853184,
+        "recall": 0.9188773593764673
+      },
+      "eval_TELEPHONENUM": {
+        "f1": 0.9914734842872345,
+        "number": 36941,
+        "precision": 0.9867285109121132,
+        "recall": 0.9962643133645543
+      },
+      "eval_ZIPCODE": {
+        "f1": 0.979114041032592,
+        "number": 24346,
+        "precision": 0.9790939335441738,
+        "recall": 0.9791341493469153
+      },
+      "eval_overall_accuracy": 0.9925094738116392,
+      "eval_overall_f1": 0.9573807470200791,
+      "eval_overall_precision": 0.9510517244513335,
+      "eval_overall_recall": 0.9637945701903046,
+      "eval_runtime": 3723.6953,
+      "eval_samples_per_second": 87.435,
+      "eval_steps_per_second": 43.718,
+      "step": 244188
+    },
+    {
+      "epoch": 3.0004422862513476,
+      "grad_norm": 0.4948091506958008,
+      "learning_rate": 1.6999570003439974e-05,
+      "loss": 0.5349,
+      "step": 244224
+    },
+    {
+      "epoch": 3.0035874329275973,
+      "grad_norm": 72.91806030273438,
+      "learning_rate": 1.699642488574377e-05,
+      "loss": 0.5571,
+      "step": 244480
+    },
+    {
+      "epoch": 3.0067325796038467,
+      "grad_norm": 4.478951930999756,
+      "learning_rate": 1.699327976804757e-05,
+      "loss": 0.5523,
+      "step": 244736
+    },
+    {
+      "epoch": 3.009877726280096,
+      "grad_norm": 16.737342834472656,
+      "learning_rate": 1.699013465035137e-05,
+      "loss": 0.4609,
+      "step": 244992
+    },
+    {
+      "epoch": 3.0130228729563457,
+      "grad_norm": 17.36398696899414,
+      "learning_rate": 1.698698953265517e-05,
+      "loss": 0.4823,
+      "step": 245248
+    },
+    {
+      "epoch": 3.016168019632595,
+      "grad_norm": 82.36427307128906,
+      "learning_rate": 1.6983844414958966e-05,
+      "loss": 0.5111,
+      "step": 245504
+    },
+    {
+      "epoch": 3.019313166308845,
+      "grad_norm": 26.42330551147461,
+      "learning_rate": 1.6980699297262766e-05,
+      "loss": 0.4467,
+      "step": 245760
+    },
+    {
+      "epoch": 3.022458312985094,
+      "grad_norm": 7.848747253417969,
+      "learning_rate": 1.6977554179566566e-05,
+      "loss": 0.5018,
+      "step": 246016
+    },
+    {
+      "epoch": 3.025603459661344,
+      "grad_norm": 2.3885860443115234,
+      "learning_rate": 1.6974409061870362e-05,
+      "loss": 0.5183,
+      "step": 246272
+    },
+    {
+      "epoch": 3.0287486063375932,
+      "grad_norm": 15.767975807189941,
+      "learning_rate": 1.6971263944174162e-05,
+      "loss": 0.5469,
+      "step": 246528
+    },
+    {
+      "epoch": 3.031893753013843,
+      "grad_norm": 16.330413818359375,
+      "learning_rate": 1.6968118826477962e-05,
+      "loss": 0.451,
+      "step": 246784
+    },
+    {
+      "epoch": 3.0350388996900923,
+      "grad_norm": 0.34589022397994995,
+      "learning_rate": 1.696497370878176e-05,
+      "loss": 0.4811,
+      "step": 247040
+    },
+    {
+      "epoch": 3.038184046366342,
+      "grad_norm": 26.566730499267578,
+      "learning_rate": 1.6961828591085558e-05,
+      "loss": 0.5117,
+      "step": 247296
+    },
+    {
+      "epoch": 3.0413291930425914,
+      "grad_norm": 0.23913775384426117,
+      "learning_rate": 1.6958683473389358e-05,
+      "loss": 0.4433,
+      "step": 247552
+    },
+    {
+      "epoch": 3.044474339718841,
+      "grad_norm": 1.7148301601409912,
+      "learning_rate": 1.6955538355693154e-05,
+      "loss": 0.469,
+      "step": 247808
+    },
+    {
+      "epoch": 3.0476194863950905,
+      "grad_norm": 0.237884521484375,
+      "learning_rate": 1.6952393237996954e-05,
+      "loss": 0.4618,
+      "step": 248064
+    },
+    {
+      "epoch": 3.0507646330713403,
+      "grad_norm": 25.270477294921875,
+      "learning_rate": 1.6949248120300754e-05,
+      "loss": 0.4415,
+      "step": 248320
+    },
+    {
+      "epoch": 3.0539097797475896,
+      "grad_norm": 598.02197265625,
+      "learning_rate": 1.6946103002604554e-05,
+      "loss": 0.4995,
+      "step": 248576
+    },
+    {
+      "epoch": 3.0570549264238394,
+      "grad_norm": 62.90739440917969,
+      "learning_rate": 1.694295788490835e-05,
+      "loss": 0.3828,
+      "step": 248832
+    },
+    {
+      "epoch": 3.0602000731000887,
+      "grad_norm": 0.3220573365688324,
+      "learning_rate": 1.693981276721215e-05,
+      "loss": 0.4903,
+      "step": 249088
+    },
+    {
+      "epoch": 3.0633452197763384,
+      "grad_norm": 6.804289817810059,
+      "learning_rate": 1.6936667649515946e-05,
+      "loss": 0.4257,
+      "step": 249344
+    },
+    {
+      "epoch": 3.0664903664525878,
+      "grad_norm": 0.06001316010951996,
+      "learning_rate": 1.6933522531819746e-05,
+      "loss": 0.5471,
+      "step": 249600
+    },
+    {
+      "epoch": 3.0696355131288375,
+      "grad_norm": 43.54471206665039,
+      "learning_rate": 1.6930377414123546e-05,
+      "loss": 0.4339,
+      "step": 249856
+    },
+    {
+      "epoch": 3.072780659805087,
+      "grad_norm": 0.04313374683260918,
+      "learning_rate": 1.6927232296427346e-05,
+      "loss": 0.5475,
+      "step": 250112
+    },
+    {
+      "epoch": 3.0759258064813366,
+      "grad_norm": 1.3913486003875732,
+      "learning_rate": 1.6924087178731142e-05,
+      "loss": 0.4765,
+      "step": 250368
+    },
+    {
+      "epoch": 3.079070953157586,
+      "grad_norm": 1.6977300643920898,
+      "learning_rate": 1.6920942061034942e-05,
+      "loss": 0.4393,
+      "step": 250624
+    },
+    {
+      "epoch": 3.0822160998338357,
+      "grad_norm": 228.96875,
+      "learning_rate": 1.691779694333874e-05,
+      "loss": 0.6406,
+      "step": 250880
+    },
+    {
+      "epoch": 3.085361246510085,
+      "grad_norm": 155.6009063720703,
+      "learning_rate": 1.6914651825642538e-05,
+      "loss": 0.4764,
+      "step": 251136
+    },
+    {
+      "epoch": 3.088506393186335,
+      "grad_norm": 0.21487553417682648,
+      "learning_rate": 1.6911506707946338e-05,
+      "loss": 0.5206,
+      "step": 251392
+    },
+    {
+      "epoch": 3.091651539862584,
+      "grad_norm": 31.710140228271484,
+      "learning_rate": 1.6908361590250138e-05,
+      "loss": 0.5498,
+      "step": 251648
+    },
+    {
+      "epoch": 3.0947966865388334,
+      "grad_norm": 18.11326026916504,
+      "learning_rate": 1.6905216472553938e-05,
+      "loss": 0.4761,
+      "step": 251904
+    },
+    {
+      "epoch": 3.097941833215083,
+      "grad_norm": 0.6825530529022217,
+      "learning_rate": 1.6902071354857734e-05,
+      "loss": 0.4681,
+      "step": 252160
+    },
+    {
+      "epoch": 3.1010869798913325,
+      "grad_norm": 0.8099691271781921,
+      "learning_rate": 1.6898926237161534e-05,
+      "loss": 0.5021,
+      "step": 252416
+    },
+    {
+      "epoch": 3.1042321265675823,
+      "grad_norm": 0.22034801542758942,
+      "learning_rate": 1.689578111946533e-05,
+      "loss": 0.5279,
+      "step": 252672
+    },
+    {
+      "epoch": 3.1073772732438316,
+      "grad_norm": 2.8035099506378174,
+      "learning_rate": 1.689263600176913e-05,
+      "loss": 0.5461,
+      "step": 252928
+    },
+    {
+      "epoch": 3.1105224199200814,
+      "grad_norm": 2.394355535507202,
+      "learning_rate": 1.6889490884072926e-05,
+      "loss": 0.4819,
+      "step": 253184
+    },
+    {
+      "epoch": 3.1136675665963307,
+      "grad_norm": 44.298828125,
+      "learning_rate": 1.688634576637673e-05,
+      "loss": 0.4995,
+      "step": 253440
+    },
+    {
+      "epoch": 3.1168127132725805,
+      "grad_norm": 11.29290771484375,
+      "learning_rate": 1.6883200648680526e-05,
+      "loss": 0.5052,
+      "step": 253696
+    },
+    {
+      "epoch": 3.11995785994883,
+      "grad_norm": 0.14054346084594727,
+      "learning_rate": 1.6880055530984326e-05,
+      "loss": 0.4331,
+      "step": 253952
+    },
+    {
+      "epoch": 3.1231030066250796,
+      "grad_norm": 65.63410186767578,
+      "learning_rate": 1.6876910413288122e-05,
+      "loss": 0.5042,
+      "step": 254208
+    },
+    {
+      "epoch": 3.126248153301329,
+      "grad_norm": 4.680383205413818,
+      "learning_rate": 1.6873765295591922e-05,
+      "loss": 0.5132,
+      "step": 254464
+    },
+    {
+      "epoch": 3.1293932999775786,
+      "grad_norm": 3.124560594558716,
+      "learning_rate": 1.6870620177895722e-05,
+      "loss": 0.4592,
+      "step": 254720
+    },
+    {
+      "epoch": 3.132538446653828,
+      "grad_norm": 1.1461992263793945,
+      "learning_rate": 1.686747506019952e-05,
+      "loss": 0.4445,
+      "step": 254976
+    },
+    {
+      "epoch": 3.1356835933300777,
+      "grad_norm": 13.09837532043457,
+      "learning_rate": 1.6864329942503318e-05,
+      "loss": 0.544,
+      "step": 255232
+    },
+    {
+      "epoch": 3.138828740006327,
+      "grad_norm": 211.1959228515625,
+      "learning_rate": 1.6861184824807118e-05,
+      "loss": 0.474,
+      "step": 255488
+    },
+    {
+      "epoch": 3.141973886682577,
+      "grad_norm": 222.2516632080078,
+      "learning_rate": 1.6858039707110914e-05,
+      "loss": 0.4765,
+      "step": 255744
+    },
+    {
+      "epoch": 3.145119033358826,
+      "grad_norm": 10.772210121154785,
+      "learning_rate": 1.6854894589414714e-05,
+      "loss": 0.5102,
+      "step": 256000
+    },
+    {
+      "epoch": 3.148264180035076,
+      "grad_norm": 4.446112155914307,
+      "learning_rate": 1.6851749471718514e-05,
+      "loss": 0.5625,
+      "step": 256256
+    },
+    {
+      "epoch": 3.1514093267113252,
+      "grad_norm": 0.4195280373096466,
+      "learning_rate": 1.684860435402231e-05,
+      "loss": 0.553,
+      "step": 256512
+    },
+    {
+      "epoch": 3.154554473387575,
+      "grad_norm": 1.3770719766616821,
+      "learning_rate": 1.684545923632611e-05,
+      "loss": 0.4071,
+      "step": 256768
+    },
+    {
+      "epoch": 3.1576996200638243,
+      "grad_norm": 0.2291870266199112,
+      "learning_rate": 1.684231411862991e-05,
+      "loss": 0.4442,
+      "step": 257024
+    },
+    {
+      "epoch": 3.160844766740074,
+      "grad_norm": 187.40231323242188,
+      "learning_rate": 1.683916900093371e-05,
+      "loss": 0.4538,
+      "step": 257280
+    },
+    {
+      "epoch": 3.1639899134163234,
+      "grad_norm": 0.19226644933223724,
+      "learning_rate": 1.6836023883237506e-05,
+      "loss": 0.5483,
+      "step": 257536
+    },
+    {
+      "epoch": 3.167135060092573,
+      "grad_norm": 0.3161170482635498,
+      "learning_rate": 1.6832878765541306e-05,
+      "loss": 0.5218,
+      "step": 257792
+    },
+    {
+      "epoch": 3.1702802067688225,
+      "grad_norm": 21.98072624206543,
+      "learning_rate": 1.6829733647845102e-05,
+      "loss": 0.4609,
+      "step": 258048
+    },
+    {
+      "epoch": 3.1734253534450723,
+      "grad_norm": 0.3211471438407898,
+      "learning_rate": 1.6826588530148906e-05,
+      "loss": 0.4235,
+      "step": 258304
+    },
+    {
+      "epoch": 3.1765705001213216,
+      "grad_norm": 197.44752502441406,
+      "learning_rate": 1.6823443412452702e-05,
+      "loss": 0.4854,
+      "step": 258560
+    },
+    {
+      "epoch": 3.179715646797571,
+      "grad_norm": 216.93283081054688,
+      "learning_rate": 1.6820298294756502e-05,
+      "loss": 0.5095,
+      "step": 258816
+    },
+    {
+      "epoch": 3.1828607934738207,
+      "grad_norm": 0.3522017002105713,
+      "learning_rate": 1.6817153177060298e-05,
+      "loss": 0.5051,
+      "step": 259072
+    },
+    {
+      "epoch": 3.18600594015007,
+      "grad_norm": 58.243167877197266,
+      "learning_rate": 1.6814008059364098e-05,
+      "loss": 0.4807,
+      "step": 259328
+    },
+    {
+      "epoch": 3.1891510868263198,
+      "grad_norm": 13.860437393188477,
+      "learning_rate": 1.6810862941667894e-05,
+      "loss": 0.5928,
+      "step": 259584
+    },
+    {
+      "epoch": 3.192296233502569,
+      "grad_norm": 62.33819580078125,
+      "learning_rate": 1.6807717823971694e-05,
+      "loss": 0.4755,
+      "step": 259840
+    },
+    {
+      "epoch": 3.195441380178819,
+      "grad_norm": 314.954833984375,
+      "learning_rate": 1.6804572706275494e-05,
+      "loss": 0.4656,
+      "step": 260096
+    },
+    {
+      "epoch": 3.198586526855068,
+      "grad_norm": 0.050347741693258286,
+      "learning_rate": 1.6801427588579294e-05,
+      "loss": 0.5019,
+      "step": 260352
+    },
+    {
+      "epoch": 3.201731673531318,
+      "grad_norm": 12.184664726257324,
+      "learning_rate": 1.679828247088309e-05,
+      "loss": 0.4917,
+      "step": 260608
+    },
+    {
+      "epoch": 3.2048768202075673,
+      "grad_norm": 0.3007303476333618,
+      "learning_rate": 1.679513735318689e-05,
+      "loss": 0.5304,
+      "step": 260864
+    },
+    {
+      "epoch": 3.208021966883817,
+      "grad_norm": 3.9271748065948486,
+      "learning_rate": 1.679199223549069e-05,
+      "loss": 0.4712,
+      "step": 261120
+    },
+    {
+      "epoch": 3.2111671135600663,
+      "grad_norm": 25.460609436035156,
+      "learning_rate": 1.6788847117794486e-05,
+      "loss": 0.6045,
+      "step": 261376
+    },
+    {
+      "epoch": 3.214312260236316,
+      "grad_norm": 465.91766357421875,
+      "learning_rate": 1.6785702000098286e-05,
+      "loss": 0.5178,
+      "step": 261632
+    },
+    {
+      "epoch": 3.2174574069125654,
+      "grad_norm": 1.2809616327285767,
+      "learning_rate": 1.6782556882402086e-05,
+      "loss": 0.5214,
+      "step": 261888
+    },
+    {
+      "epoch": 3.220602553588815,
+      "grad_norm": 1.410000205039978,
+      "learning_rate": 1.6779411764705886e-05,
+      "loss": 0.5069,
+      "step": 262144
+    },
+    {
+      "epoch": 3.2237477002650645,
+      "grad_norm": 1.05242919921875,
+      "learning_rate": 1.6776266647009682e-05,
+      "loss": 0.5353,
+      "step": 262400
+    },
+    {
+      "epoch": 3.2268928469413143,
+      "grad_norm": 2.205047607421875,
+      "learning_rate": 1.6773121529313482e-05,
+      "loss": 0.611,
+      "step": 262656
+    },
+    {
+      "epoch": 3.2300379936175636,
+      "grad_norm": 0.5542662739753723,
+      "learning_rate": 1.6769976411617278e-05,
+      "loss": 0.4971,
+      "step": 262912
+    },
+    {
+      "epoch": 3.2331831402938134,
+      "grad_norm": 182.70164489746094,
+      "learning_rate": 1.6766831293921078e-05,
+      "loss": 0.4656,
+      "step": 263168
+    },
+    {
+      "epoch": 3.2363282869700627,
+      "grad_norm": 0.3895760178565979,
+      "learning_rate": 1.6763686176224878e-05,
+      "loss": 0.5441,
+      "step": 263424
+    },
+    {
+      "epoch": 3.2394734336463125,
+      "grad_norm": 1.8494099378585815,
+      "learning_rate": 1.6760541058528678e-05,
+      "loss": 0.494,
+      "step": 263680
+    },
+    {
+      "epoch": 3.242618580322562,
+      "grad_norm": 108.67359161376953,
+      "learning_rate": 1.6757395940832474e-05,
+      "loss": 0.5363,
+      "step": 263936
+    },
+    {
+      "epoch": 3.2457637269988115,
+      "grad_norm": 0.44472789764404297,
+      "learning_rate": 1.6754250823136274e-05,
+      "loss": 0.5913,
+      "step": 264192
+    },
+    {
+      "epoch": 3.248908873675061,
+      "grad_norm": 94.2912368774414,
+      "learning_rate": 1.675110570544007e-05,
+      "loss": 0.5095,
+      "step": 264448
+    },
+    {
+      "epoch": 3.25205402035131,
+      "grad_norm": 20.682180404663086,
+      "learning_rate": 1.674796058774387e-05,
+      "loss": 0.4988,
+      "step": 264704
+    },
+    {
+      "epoch": 3.25519916702756,
+      "grad_norm": 0.455321341753006,
+      "learning_rate": 1.674481547004767e-05,
+      "loss": 0.5389,
+      "step": 264960
+    },
+    {
+      "epoch": 3.2583443137038097,
+      "grad_norm": 34.81191635131836,
+      "learning_rate": 1.674167035235147e-05,
+      "loss": 0.5167,
+      "step": 265216
+    },
+    {
+      "epoch": 3.261489460380059,
+      "grad_norm": 1.3689322471618652,
+      "learning_rate": 1.6738525234655266e-05,
+      "loss": 0.4903,
+      "step": 265472
+    },
+    {
+      "epoch": 3.2646346070563084,
+      "grad_norm": 53.11245346069336,
+      "learning_rate": 1.6735380116959066e-05,
+      "loss": 0.4854,
+      "step": 265728
+    },
+    {
+      "epoch": 3.267779753732558,
+      "grad_norm": 3.8486721515655518,
+      "learning_rate": 1.6732234999262866e-05,
+      "loss": 0.4486,
+      "step": 265984
+    },
+    {
+      "epoch": 3.2709249004088075,
+      "grad_norm": 1.543716311454773,
+      "learning_rate": 1.6729089881566662e-05,
+      "loss": 0.4874,
+      "step": 266240
+    },
+    {
+      "epoch": 3.2740700470850572,
+      "grad_norm": 0.47223401069641113,
+      "learning_rate": 1.6725944763870462e-05,
+      "loss": 0.4081,
+      "step": 266496
+    },
+    {
+      "epoch": 3.2772151937613065,
+      "grad_norm": 18.2301025390625,
+      "learning_rate": 1.6722799646174262e-05,
+      "loss": 0.5056,
+      "step": 266752
+    },
+    {
+      "epoch": 3.2803603404375563,
+      "grad_norm": 0.7127645015716553,
+      "learning_rate": 1.6719654528478058e-05,
+      "loss": 0.5375,
+      "step": 267008
+    },
+    {
+      "epoch": 3.2835054871138056,
+      "grad_norm": 0.5302969813346863,
+      "learning_rate": 1.6716509410781858e-05,
+      "loss": 0.5051,
+      "step": 267264
+    },
+    {
+      "epoch": 3.2866506337900554,
+      "grad_norm": 23.452619552612305,
+      "learning_rate": 1.6713364293085658e-05,
+      "loss": 0.4686,
+      "step": 267520
+    },
+    {
+      "epoch": 3.2897957804663047,
+      "grad_norm": 95.13858795166016,
+      "learning_rate": 1.6710219175389454e-05,
+      "loss": 0.5505,
+      "step": 267776
+    },
+    {
+      "epoch": 3.2929409271425545,
+      "grad_norm": 231.501708984375,
+      "learning_rate": 1.6707074057693254e-05,
+      "loss": 0.5001,
+      "step": 268032
+    },
+    {
+      "epoch": 3.296086073818804,
+      "grad_norm": 3.7133214473724365,
+      "learning_rate": 1.6703928939997054e-05,
+      "loss": 0.4919,
+      "step": 268288
+    },
+    {
+      "epoch": 3.2992312204950536,
+      "grad_norm": 6.527437686920166,
+      "learning_rate": 1.6700783822300854e-05,
+      "loss": 0.5303,
+      "step": 268544
+    },
+    {
+      "epoch": 3.302376367171303,
+      "grad_norm": 0.9041167497634888,
+      "learning_rate": 1.669763870460465e-05,
+      "loss": 0.5724,
+      "step": 268800
+    },
+    {
+      "epoch": 3.3055215138475527,
+      "grad_norm": 37.12371826171875,
+      "learning_rate": 1.669449358690845e-05,
+      "loss": 0.5259,
+      "step": 269056
+    },
+    {
+      "epoch": 3.308666660523802,
+      "grad_norm": 1.0556756258010864,
+      "learning_rate": 1.6691348469212246e-05,
+      "loss": 0.4594,
+      "step": 269312
+    },
+    {
+      "epoch": 3.3118118072000517,
+      "grad_norm": 0.1722612977027893,
+      "learning_rate": 1.6688203351516046e-05,
+      "loss": 0.573,
+      "step": 269568
+    },
+    {
+      "epoch": 3.314956953876301,
+      "grad_norm": 1.3218982219696045,
+      "learning_rate": 1.6685058233819842e-05,
+      "loss": 0.4533,
+      "step": 269824
+    },
+    {
+      "epoch": 3.318102100552551,
+      "grad_norm": 8.506532669067383,
+      "learning_rate": 1.6681913116123646e-05,
+      "loss": 0.4492,
+      "step": 270080
+    },
+    {
+      "epoch": 3.3212472472288,
+      "grad_norm": 37.350807189941406,
+      "learning_rate": 1.6678767998427442e-05,
+      "loss": 0.5392,
+      "step": 270336
+    },
+    {
+      "epoch": 3.32439239390505,
+      "grad_norm": 24.144933700561523,
+      "learning_rate": 1.6675622880731242e-05,
+      "loss": 0.565,
+      "step": 270592
+    },
+    {
+      "epoch": 3.3275375405812992,
+      "grad_norm": 7.205847263336182,
+      "learning_rate": 1.667247776303504e-05,
+      "loss": 0.4834,
+      "step": 270848
+    },
+    {
+      "epoch": 3.330682687257549,
+      "grad_norm": 0.3731878995895386,
+      "learning_rate": 1.6669332645338838e-05,
+      "loss": 0.5598,
+      "step": 271104
+    },
+    {
+      "epoch": 3.3338278339337983,
+      "grad_norm": 0.8721461892127991,
+      "learning_rate": 1.6666187527642638e-05,
+      "loss": 0.4995,
+      "step": 271360
+    },
+    {
+      "epoch": 3.3369729806100477,
+      "grad_norm": 106.60723876953125,
+      "learning_rate": 1.6663042409946438e-05,
+      "loss": 0.5368,
+      "step": 271616
+    },
+    {
+      "epoch": 3.3401181272862974,
+      "grad_norm": 112.32515716552734,
+      "learning_rate": 1.6659897292250234e-05,
+      "loss": 0.4475,
+      "step": 271872
+    },
+    {
+      "epoch": 3.343263273962547,
+      "grad_norm": 121.46126556396484,
+      "learning_rate": 1.6656752174554034e-05,
+      "loss": 0.5106,
+      "step": 272128
+    },
+    {
+      "epoch": 3.3464084206387965,
+      "grad_norm": 13.887198448181152,
+      "learning_rate": 1.6653607056857834e-05,
+      "loss": 0.4732,
+      "step": 272384
+    },
+    {
+      "epoch": 3.349553567315046,
+      "grad_norm": 0.08012892305850983,
+      "learning_rate": 1.665046193916163e-05,
+      "loss": 0.552,
+      "step": 272640
+    },
+    {
+      "epoch": 3.3526987139912956,
+      "grad_norm": 1.2425813674926758,
+      "learning_rate": 1.664731682146543e-05,
+      "loss": 0.4831,
+      "step": 272896
+    },
+    {
+      "epoch": 3.355843860667545,
+      "grad_norm": 1.8234463930130005,
+      "learning_rate": 1.6644171703769226e-05,
+      "loss": 0.507,
+      "step": 273152
+    },
+    {
+      "epoch": 3.3589890073437947,
+      "grad_norm": 0.9249053597450256,
+      "learning_rate": 1.664102658607303e-05,
+      "loss": 0.5295,
+      "step": 273408
+    },
+    {
+      "epoch": 3.362134154020044,
+      "grad_norm": 0.9175606369972229,
+      "learning_rate": 1.6637881468376826e-05,
+      "loss": 0.4655,
+      "step": 273664
+    },
+    {
+      "epoch": 3.3652793006962938,
+      "grad_norm": 154.72328186035156,
+      "learning_rate": 1.6634736350680626e-05,
+      "loss": 0.5649,
+      "step": 273920
+    },
+    {
+      "epoch": 3.368424447372543,
+      "grad_norm": 90.13003540039062,
+      "learning_rate": 1.6631591232984422e-05,
+      "loss": 0.6428,
+      "step": 274176
+    },
+    {
+      "epoch": 3.371569594048793,
+      "grad_norm": 55.9456672668457,
+      "learning_rate": 1.6628446115288222e-05,
+      "loss": 0.5794,
+      "step": 274432
+    },
+    {
+      "epoch": 3.374714740725042,
+      "grad_norm": 34.19145965576172,
+      "learning_rate": 1.662530099759202e-05,
+      "loss": 0.4947,
+      "step": 274688
+    },
+    {
+      "epoch": 3.377859887401292,
+      "grad_norm": 0.03633107244968414,
+      "learning_rate": 1.6622155879895818e-05,
+      "loss": 0.5793,
+      "step": 274944
+    },
+    {
+      "epoch": 3.3810050340775413,
+      "grad_norm": 60.763710021972656,
+      "learning_rate": 1.6619010762199618e-05,
+      "loss": 0.5055,
+      "step": 275200
+    },
+    {
+      "epoch": 3.384150180753791,
+      "grad_norm": 3.0623834133148193,
+      "learning_rate": 1.6615865644503418e-05,
+      "loss": 0.4545,
+      "step": 275456
+    },
+    {
+      "epoch": 3.3872953274300404,
+      "grad_norm": 0.1399703025817871,
+      "learning_rate": 1.6612720526807214e-05,
+      "loss": 0.5938,
+      "step": 275712
+    },
+    {
+      "epoch": 3.39044047410629,
+      "grad_norm": 0.4174153804779053,
+      "learning_rate": 1.6609575409111014e-05,
+      "loss": 0.5943,
+      "step": 275968
+    },
+    {
+      "epoch": 3.3935856207825394,
+      "grad_norm": 21.17488670349121,
+      "learning_rate": 1.6606430291414814e-05,
+      "loss": 0.4772,
+      "step": 276224
+    },
+    {
+      "epoch": 3.396730767458789,
+      "grad_norm": 1.6206845045089722,
+      "learning_rate": 1.660328517371861e-05,
+      "loss": 0.5516,
+      "step": 276480
+    },
+    {
+      "epoch": 3.3998759141350385,
+      "grad_norm": 0.24734431505203247,
+      "learning_rate": 1.660014005602241e-05,
+      "loss": 0.4474,
+      "step": 276736
+    },
+    {
+      "epoch": 3.4030210608112883,
+      "grad_norm": 48.65492248535156,
+      "learning_rate": 1.659699493832621e-05,
+      "loss": 0.5816,
+      "step": 276992
+    },
+    {
+      "epoch": 3.4061662074875376,
+      "grad_norm": 2.6201436519622803,
+      "learning_rate": 1.659384982063001e-05,
+      "loss": 0.4028,
+      "step": 277248
+    },
+    {
+      "epoch": 3.4093113541637874,
+      "grad_norm": 53.09165954589844,
+      "learning_rate": 1.6590704702933806e-05,
+      "loss": 0.5351,
+      "step": 277504
+    },
+    {
+      "epoch": 3.4124565008400367,
+      "grad_norm": 10.02964973449707,
+      "learning_rate": 1.6587559585237606e-05,
+      "loss": 0.4621,
+      "step": 277760
+    },
+    {
+      "epoch": 3.4156016475162865,
+      "grad_norm": 374.3504943847656,
+      "learning_rate": 1.6584414467541402e-05,
+      "loss": 0.4352,
+      "step": 278016
+    },
+    {
+      "epoch": 3.418746794192536,
+      "grad_norm": 51.78233337402344,
+      "learning_rate": 1.6581269349845202e-05,
+      "loss": 0.5841,
+      "step": 278272
+    },
+    {
+      "epoch": 3.421891940868785,
+      "grad_norm": 5.724061489105225,
+      "learning_rate": 1.6578124232149002e-05,
+      "loss": 0.5283,
+      "step": 278528
+    },
+    {
+      "epoch": 3.425037087545035,
+      "grad_norm": 0.0579860582947731,
+      "learning_rate": 1.6574979114452802e-05,
+      "loss": 0.5386,
+      "step": 278784
+    },
+    {
+      "epoch": 3.4281822342212847,
+      "grad_norm": 21.401683807373047,
+      "learning_rate": 1.6571833996756598e-05,
+      "loss": 0.5005,
+      "step": 279040
+    },
+    {
+      "epoch": 3.431327380897534,
+      "grad_norm": 3.490536689758301,
+      "learning_rate": 1.6568688879060398e-05,
+      "loss": 0.5714,
+      "step": 279296
+    },
+    {
+      "epoch": 3.4344725275737833,
+      "grad_norm": 44.731632232666016,
+      "learning_rate": 1.6565543761364194e-05,
+      "loss": 0.4596,
+      "step": 279552
+    },
+    {
+      "epoch": 3.437617674250033,
+      "grad_norm": 1.0921201705932617,
+      "learning_rate": 1.6562398643667994e-05,
+      "loss": 0.5102,
+      "step": 279808
+    },
+    {
+      "epoch": 3.4407628209262824,
+      "grad_norm": 13.510580062866211,
+      "learning_rate": 1.6559253525971794e-05,
+      "loss": 0.5003,
+      "step": 280064
+    },
+    {
+      "epoch": 3.443907967602532,
+      "grad_norm": 1.5561238527297974,
+      "learning_rate": 1.6556108408275594e-05,
+      "loss": 0.4669,
+      "step": 280320
+    },
+    {
+      "epoch": 3.4470531142787815,
+      "grad_norm": 33.022396087646484,
+      "learning_rate": 1.655296329057939e-05,
+      "loss": 0.563,
+      "step": 280576
+    },
+    {
+      "epoch": 3.4501982609550312,
+      "grad_norm": 1.3516336679458618,
+      "learning_rate": 1.654981817288319e-05,
+      "loss": 0.4832,
+      "step": 280832
+    },
+    {
+      "epoch": 3.4533434076312806,
+      "grad_norm": 60.90022277832031,
+      "learning_rate": 1.6546673055186986e-05,
+      "loss": 0.5817,
+      "step": 281088
+    },
+    {
+      "epoch": 3.4564885543075303,
+      "grad_norm": 15.441397666931152,
+      "learning_rate": 1.6543527937490786e-05,
+      "loss": 0.5002,
+      "step": 281344
+    },
+    {
+      "epoch": 3.4596337009837796,
+      "grad_norm": 22.145368576049805,
+      "learning_rate": 1.6540382819794586e-05,
+      "loss": 0.4889,
+      "step": 281600
+    },
+    {
+      "epoch": 3.4627788476600294,
+      "grad_norm": 2.654942750930786,
+      "learning_rate": 1.6537237702098386e-05,
+      "loss": 0.478,
+      "step": 281856
+    },
+    {
+      "epoch": 3.4659239943362787,
+      "grad_norm": 16.67755126953125,
+      "learning_rate": 1.6534092584402182e-05,
+      "loss": 0.5456,
+      "step": 282112
+    },
+    {
+      "epoch": 3.4690691410125285,
+      "grad_norm": 97.49522399902344,
+      "learning_rate": 1.6530947466705982e-05,
+      "loss": 0.4822,
+      "step": 282368
+    },
+    {
+      "epoch": 3.472214287688778,
+      "grad_norm": 10.756332397460938,
+      "learning_rate": 1.6527802349009782e-05,
+      "loss": 0.4693,
+      "step": 282624
+    },
+    {
+      "epoch": 3.4753594343650276,
+      "grad_norm": 2.344128131866455,
+      "learning_rate": 1.6524657231313578e-05,
+      "loss": 0.4803,
+      "step": 282880
+    },
+    {
+      "epoch": 3.478504581041277,
+      "grad_norm": 0.4296514391899109,
+      "learning_rate": 1.6521512113617378e-05,
+      "loss": 0.539,
+      "step": 283136
+    },
+    {
+      "epoch": 3.4816497277175267,
+      "grad_norm": 158.57447814941406,
+      "learning_rate": 1.6518366995921178e-05,
+      "loss": 0.4677,
+      "step": 283392
+    },
+    {
+      "epoch": 3.484794874393776,
+      "grad_norm": 4.016801357269287,
+      "learning_rate": 1.6515221878224978e-05,
+      "loss": 0.4684,
+      "step": 283648
+    },
+    {
+      "epoch": 3.4879400210700258,
+      "grad_norm": 0.4205678403377533,
+      "learning_rate": 1.6512076760528774e-05,
+      "loss": 0.5426,
+      "step": 283904
+    },
+    {
+      "epoch": 3.491085167746275,
+      "grad_norm": 66.58230590820312,
+      "learning_rate": 1.6508931642832574e-05,
+      "loss": 0.4711,
+      "step": 284160
+    },
+    {
+      "epoch": 3.4942303144225244,
+      "grad_norm": 8.684982299804688,
+      "learning_rate": 1.650578652513637e-05,
+      "loss": 0.4695,
+      "step": 284416
+    },
+    {
+      "epoch": 3.497375461098774,
+      "grad_norm": 0.9667931199073792,
+      "learning_rate": 1.650264140744017e-05,
+      "loss": 0.5183,
+      "step": 284672
+    },
+    {
+      "epoch": 3.500520607775024,
+      "grad_norm": 59.399803161621094,
+      "learning_rate": 1.6499496289743967e-05,
+      "loss": 0.5685,
+      "step": 284928
+    },
+    {
+      "epoch": 3.5036657544512733,
+      "grad_norm": 2.334911584854126,
+      "learning_rate": 1.649635117204777e-05,
+      "loss": 0.4244,
+      "step": 285184
+    },
+    {
+      "epoch": 3.5068109011275226,
+      "grad_norm": 0.647039532661438,
+      "learning_rate": 1.6493206054351566e-05,
+      "loss": 0.4933,
+      "step": 285440
+    },
+    {
+      "epoch": 3.5099560478037723,
+      "grad_norm": 0.03905763849616051,
+      "learning_rate": 1.6490060936655366e-05,
+      "loss": 0.4837,
+      "step": 285696
+    },
+    {
+      "epoch": 3.513101194480022,
+      "grad_norm": 15.193082809448242,
+      "learning_rate": 1.6486915818959162e-05,
+      "loss": 0.4253,
+      "step": 285952
+    },
+    {
+      "epoch": 3.5162463411562714,
+      "grad_norm": 61.416297912597656,
+      "learning_rate": 1.6483770701262962e-05,
+      "loss": 0.5753,
+      "step": 286208
+    },
+    {
+      "epoch": 3.5193914878325208,
+      "grad_norm": 0.10716848820447922,
+      "learning_rate": 1.6480625583566762e-05,
+      "loss": 0.4491,
+      "step": 286464
+    },
+    {
+      "epoch": 3.5225366345087705,
+      "grad_norm": 0.786285400390625,
+      "learning_rate": 1.6477480465870562e-05,
+      "loss": 0.5287,
+      "step": 286720
+    },
+    {
+      "epoch": 3.52568178118502,
+      "grad_norm": 0.08423032611608505,
+      "learning_rate": 1.6474335348174358e-05,
+      "loss": 0.4975,
+      "step": 286976
+    },
+    {
+      "epoch": 3.5288269278612696,
+      "grad_norm": 135.55809020996094,
+      "learning_rate": 1.6471190230478158e-05,
+      "loss": 0.5153,
+      "step": 287232
+    },
+    {
+      "epoch": 3.531972074537519,
+      "grad_norm": 421.2353820800781,
+      "learning_rate": 1.6468045112781958e-05,
+      "loss": 0.5046,
+      "step": 287488
+    },
+    {
+      "epoch": 3.5351172212137687,
+      "grad_norm": 267.75115966796875,
+      "learning_rate": 1.6464899995085754e-05,
+      "loss": 0.4901,
+      "step": 287744
+    },
+    {
+      "epoch": 3.538262367890018,
+      "grad_norm": 39.2845573425293,
+      "learning_rate": 1.6461754877389554e-05,
+      "loss": 0.5153,
+      "step": 288000
+    },
+    {
+      "epoch": 3.541407514566268,
+      "grad_norm": 20.747488021850586,
+      "learning_rate": 1.645860975969335e-05,
+      "loss": 0.4516,
+      "step": 288256
+    },
+    {
+      "epoch": 3.544552661242517,
+      "grad_norm": 61.285682678222656,
+      "learning_rate": 1.6455464641997154e-05,
+      "loss": 0.5375,
+      "step": 288512
+    },
+    {
+      "epoch": 3.547697807918767,
+      "grad_norm": 0.16573591530323029,
+      "learning_rate": 1.645231952430095e-05,
+      "loss": 0.5854,
+      "step": 288768
+    },
+    {
+      "epoch": 3.550842954595016,
+      "grad_norm": 0.1327620893716812,
+      "learning_rate": 1.644917440660475e-05,
+      "loss": 0.5168,
+      "step": 289024
+    },
+    {
+      "epoch": 3.553988101271266,
+      "grad_norm": 6.821236610412598,
+      "learning_rate": 1.6446029288908546e-05,
+      "loss": 0.5404,
+      "step": 289280
+    },
+    {
+      "epoch": 3.5571332479475153,
+      "grad_norm": 102.42604064941406,
+      "learning_rate": 1.6442884171212346e-05,
+      "loss": 0.4493,
+      "step": 289536
+    },
+    {
+      "epoch": 3.560278394623765,
+      "grad_norm": 0.32021015882492065,
+      "learning_rate": 1.6439739053516143e-05,
+      "loss": 0.5027,
+      "step": 289792
+    },
+    {
+      "epoch": 3.5634235413000144,
+      "grad_norm": 0.23085713386535645,
+      "learning_rate": 1.6436593935819946e-05,
+      "loss": 0.5332,
+      "step": 290048
+    },
+    {
+      "epoch": 3.5665686879762637,
+      "grad_norm": 30.993846893310547,
+      "learning_rate": 1.6433448818123742e-05,
+      "loss": 0.491,
+      "step": 290304
+    },
+    {
+      "epoch": 3.5697138346525135,
+      "grad_norm": 0.5032532811164856,
+      "learning_rate": 1.6430303700427542e-05,
+      "loss": 0.5545,
+      "step": 290560
+    },
+    {
+      "epoch": 3.5728589813287632,
+      "grad_norm": 63.583194732666016,
+      "learning_rate": 1.642715858273134e-05,
+      "loss": 0.5404,
+      "step": 290816
+    },
+    {
+      "epoch": 3.5760041280050126,
+      "grad_norm": 120.06217193603516,
+      "learning_rate": 1.6424013465035138e-05,
+      "loss": 0.5228,
+      "step": 291072
+    },
+    {
+      "epoch": 3.579149274681262,
+      "grad_norm": 22.292163848876953,
+      "learning_rate": 1.6420868347338938e-05,
+      "loss": 0.5126,
+      "step": 291328
+    },
+    {
+      "epoch": 3.5822944213575116,
+      "grad_norm": 22.050294876098633,
+      "learning_rate": 1.6417723229642734e-05,
+      "loss": 0.5669,
+      "step": 291584
+    },
+    {
+      "epoch": 3.5854395680337614,
+      "grad_norm": 1.5287801027297974,
+      "learning_rate": 1.6414578111946534e-05,
+      "loss": 0.5547,
+      "step": 291840
+    },
+    {
+      "epoch": 3.5885847147100107,
+      "grad_norm": 0.12451621145009995,
+      "learning_rate": 1.6411432994250334e-05,
+      "loss": 0.4333,
+      "step": 292096
+    },
+    {
+      "epoch": 3.59172986138626,
+      "grad_norm": 165.6510467529297,
+      "learning_rate": 1.640828787655413e-05,
+      "loss": 0.4767,
+      "step": 292352
+    },
+    {
+      "epoch": 3.59487500806251,
+      "grad_norm": 56.62143325805664,
+      "learning_rate": 1.640514275885793e-05,
+      "loss": 0.4681,
+      "step": 292608
+    },
+    {
+      "epoch": 3.5980201547387596,
+      "grad_norm": 82.96979522705078,
+      "learning_rate": 1.640199764116173e-05,
+      "loss": 0.4263,
+      "step": 292864
+    },
+    {
+      "epoch": 3.601165301415009,
+      "grad_norm": 3.0660316944122314,
+      "learning_rate": 1.6398852523465526e-05,
+      "loss": 0.526,
+      "step": 293120
+    },
+    {
+      "epoch": 3.6043104480912582,
+      "grad_norm": 332.16644287109375,
+      "learning_rate": 1.6395707405769326e-05,
+      "loss": 0.5333,
+      "step": 293376
+    },
+    {
+      "epoch": 3.607455594767508,
+      "grad_norm": 10.242597579956055,
+      "learning_rate": 1.6392562288073126e-05,
+      "loss": 0.4618,
+      "step": 293632
+    },
+    {
+      "epoch": 3.6106007414437573,
+      "grad_norm": 148.5627899169922,
+      "learning_rate": 1.6389417170376926e-05,
+      "loss": 0.6146,
+      "step": 293888
+    },
+    {
+      "epoch": 3.613745888120007,
+      "grad_norm": 13.564252853393555,
+      "learning_rate": 1.6386272052680722e-05,
+      "loss": 0.5225,
+      "step": 294144
+    },
+    {
+      "epoch": 3.6168910347962564,
+      "grad_norm": 42.36850357055664,
+      "learning_rate": 1.6383126934984522e-05,
+      "loss": 0.5449,
+      "step": 294400
+    },
+    {
+      "epoch": 3.620036181472506,
+      "grad_norm": 1.373180627822876,
+      "learning_rate": 1.637998181728832e-05,
+      "loss": 0.4743,
+      "step": 294656
+    },
+    {
+      "epoch": 3.6231813281487555,
+      "grad_norm": 27.242643356323242,
+      "learning_rate": 1.6376836699592118e-05,
+      "loss": 0.4707,
+      "step": 294912
+    },
+    {
+      "epoch": 3.6263264748250053,
+      "grad_norm": 0.23503144085407257,
+      "learning_rate": 1.6373691581895918e-05,
+      "loss": 0.583,
+      "step": 295168
+    },
+    {
+      "epoch": 3.6294716215012546,
+      "grad_norm": 17.4699649810791,
+      "learning_rate": 1.6370546464199718e-05,
+      "loss": 0.5128,
+      "step": 295424
+    },
+    {
+      "epoch": 3.6326167681775043,
+      "grad_norm": 3.355201482772827,
+      "learning_rate": 1.6367401346503514e-05,
+      "loss": 0.5238,
+      "step": 295680
+    },
+    {
+      "epoch": 3.6357619148537537,
+      "grad_norm": 37.129180908203125,
+      "learning_rate": 1.6364256228807314e-05,
+      "loss": 0.5222,
+      "step": 295936
+    },
+    {
+      "epoch": 3.6389070615300034,
+      "grad_norm": 0.04358728975057602,
+      "learning_rate": 1.636111111111111e-05,
+      "loss": 0.5514,
+      "step": 296192
+    },
+    {
+      "epoch": 3.6420522082062528,
+      "grad_norm": 6.604242324829102,
+      "learning_rate": 1.635796599341491e-05,
+      "loss": 0.4942,
+      "step": 296448
+    },
+    {
+      "epoch": 3.6451973548825025,
+      "grad_norm": 13.806446075439453,
+      "learning_rate": 1.635482087571871e-05,
+      "loss": 0.5384,
+      "step": 296704
+    },
+    {
+      "epoch": 3.648342501558752,
+      "grad_norm": 19.144948959350586,
+      "learning_rate": 1.635167575802251e-05,
+      "loss": 0.4876,
+      "step": 296960
+    },
+    {
+      "epoch": 3.651487648235001,
+      "grad_norm": 0.0893256738781929,
+      "learning_rate": 1.6348530640326306e-05,
+      "loss": 0.478,
+      "step": 297216
+    },
+    {
+      "epoch": 3.654632794911251,
+      "grad_norm": 2.3621015548706055,
+      "learning_rate": 1.6345385522630106e-05,
+      "loss": 0.5909,
+      "step": 297472
+    },
+    {
+      "epoch": 3.6577779415875007,
+      "grad_norm": 3.2733564376831055,
+      "learning_rate": 1.6342240404933906e-05,
+      "loss": 0.4492,
+      "step": 297728
+    },
+    {
+      "epoch": 3.66092308826375,
+      "grad_norm": 0.001855314476415515,
+      "learning_rate": 1.6339095287237702e-05,
+      "loss": 0.5061,
+      "step": 297984
+    },
+    {
+      "epoch": 3.6640682349399993,
+      "grad_norm": 110.29045104980469,
+      "learning_rate": 1.6335950169541502e-05,
+      "loss": 0.4995,
+      "step": 298240
+    },
+    {
+      "epoch": 3.667213381616249,
+      "grad_norm": 1.9425102472305298,
+      "learning_rate": 1.6332805051845302e-05,
+      "loss": 0.4326,
+      "step": 298496
+    },
+    {
+      "epoch": 3.670358528292499,
+      "grad_norm": 0.5388185977935791,
+      "learning_rate": 1.6329659934149102e-05,
+      "loss": 0.495,
+      "step": 298752
+    },
+    {
+      "epoch": 3.673503674968748,
+      "grad_norm": 37.57967758178711,
+      "learning_rate": 1.6326514816452898e-05,
+      "loss": 0.4679,
+      "step": 299008
+    },
+    {
+      "epoch": 3.6766488216449975,
+      "grad_norm": 254.44651794433594,
+      "learning_rate": 1.6323369698756698e-05,
+      "loss": 0.4724,
+      "step": 299264
+    },
+    {
+      "epoch": 3.6797939683212473,
+      "grad_norm": 0.08431895077228546,
+      "learning_rate": 1.6320224581060494e-05,
+      "loss": 0.4746,
+      "step": 299520
+    },
+    {
+      "epoch": 3.682939114997497,
+      "grad_norm": 60.55018997192383,
+      "learning_rate": 1.6317079463364294e-05,
+      "loss": 0.4852,
+      "step": 299776
+    },
+    {
+      "epoch": 3.6860842616737464,
+      "grad_norm": 0.30055662989616394,
+      "learning_rate": 1.631393434566809e-05,
+      "loss": 0.4301,
+      "step": 300032
+    },
+    {
+      "epoch": 3.6892294083499957,
+      "grad_norm": 55.880165100097656,
+      "learning_rate": 1.6310789227971894e-05,
+      "loss": 0.5526,
+      "step": 300288
+    },
+    {
+      "epoch": 3.6923745550262455,
+      "grad_norm": 186.46678161621094,
+      "learning_rate": 1.630764411027569e-05,
+      "loss": 0.4412,
+      "step": 300544
+    },
+    {
+      "epoch": 3.6955197017024948,
+      "grad_norm": 21.55492401123047,
+      "learning_rate": 1.630449899257949e-05,
+      "loss": 0.6048,
+      "step": 300800
+    },
+    {
+      "epoch": 3.6986648483787445,
+      "grad_norm": 0.04537157714366913,
+      "learning_rate": 1.6301353874883286e-05,
+      "loss": 0.5945,
+      "step": 301056
+    },
+    {
+      "epoch": 3.701809995054994,
+      "grad_norm": 117.09156036376953,
+      "learning_rate": 1.6298208757187086e-05,
+      "loss": 0.5149,
+      "step": 301312
+    },
+    {
+      "epoch": 3.7049551417312436,
+      "grad_norm": 19.559537887573242,
+      "learning_rate": 1.6295063639490886e-05,
+      "loss": 0.4477,
+      "step": 301568
+    },
+    {
+      "epoch": 3.708100288407493,
+      "grad_norm": 9.602081298828125,
+      "learning_rate": 1.6291918521794686e-05,
+      "loss": 0.5715,
+      "step": 301824
+    },
+    {
+      "epoch": 3.7112454350837427,
+      "grad_norm": 431.27642822265625,
+      "learning_rate": 1.6288773404098482e-05,
+      "loss": 0.4831,
+      "step": 302080
+    },
+    {
+      "epoch": 3.714390581759992,
+      "grad_norm": 89.97123718261719,
+      "learning_rate": 1.6285628286402282e-05,
+      "loss": 0.4962,
+      "step": 302336
+    },
+    {
+      "epoch": 3.717535728436242,
+      "grad_norm": 125.58710479736328,
+      "learning_rate": 1.6282483168706082e-05,
+      "loss": 0.4974,
+      "step": 302592
+    },
+    {
+      "epoch": 3.720680875112491,
+      "grad_norm": 1.5968273878097534,
+      "learning_rate": 1.627933805100988e-05,
+      "loss": 0.5484,
+      "step": 302848
+    },
+    {
+      "epoch": 3.723826021788741,
+      "grad_norm": 2.0706241130828857,
+      "learning_rate": 1.6276192933313678e-05,
+      "loss": 0.5244,
+      "step": 303104
+    },
+    {
+      "epoch": 3.72697116846499,
+      "grad_norm": 144.3460235595703,
+      "learning_rate": 1.6273047815617475e-05,
+      "loss": 0.5596,
+      "step": 303360
+    },
+    {
+      "epoch": 3.73011631514124,
+      "grad_norm": 105.07582092285156,
+      "learning_rate": 1.6269902697921274e-05,
+      "loss": 0.5752,
+      "step": 303616
+    },
+    {
+      "epoch": 3.7332614618174893,
+      "grad_norm": 0.6541815996170044,
+      "learning_rate": 1.6266757580225074e-05,
+      "loss": 0.5939,
+      "step": 303872
+    },
+    {
+      "epoch": 3.7364066084937386,
+      "grad_norm": 47.60111999511719,
+      "learning_rate": 1.6263612462528874e-05,
+      "loss": 0.4619,
+      "step": 304128
+    },
+    {
+      "epoch": 3.7395517551699884,
+      "grad_norm": 4.194868087768555,
+      "learning_rate": 1.626046734483267e-05,
+      "loss": 0.4936,
+      "step": 304384
+    },
+    {
+      "epoch": 3.742696901846238,
+      "grad_norm": 5.030226230621338,
+      "learning_rate": 1.625732222713647e-05,
+      "loss": 0.512,
+      "step": 304640
+    },
+    {
+      "epoch": 3.7458420485224875,
+      "grad_norm": 301.7357482910156,
+      "learning_rate": 1.6254177109440267e-05,
+      "loss": 0.5107,
+      "step": 304896
+    },
+    {
+      "epoch": 3.748987195198737,
+      "grad_norm": 88.85123443603516,
+      "learning_rate": 1.625103199174407e-05,
+      "loss": 0.5189,
+      "step": 305152
+    },
+    {
+      "epoch": 3.7521323418749866,
+      "grad_norm": 6.629373073577881,
+      "learning_rate": 1.6247886874047866e-05,
+      "loss": 0.5304,
+      "step": 305408
+    },
+    {
+      "epoch": 3.7552774885512363,
+      "grad_norm": 15.015730857849121,
+      "learning_rate": 1.6244741756351666e-05,
+      "loss": 0.5075,
+      "step": 305664
+    },
+    {
+      "epoch": 3.7584226352274857,
+      "grad_norm": 1.771833062171936,
+      "learning_rate": 1.6241596638655462e-05,
+      "loss": 0.4686,
+      "step": 305920
+    },
+    {
+      "epoch": 3.761567781903735,
+      "grad_norm": 87.13249206542969,
+      "learning_rate": 1.6238451520959262e-05,
+      "loss": 0.4832,
+      "step": 306176
+    },
+    {
+      "epoch": 3.7647129285799847,
+      "grad_norm": 81.09495544433594,
+      "learning_rate": 1.623530640326306e-05,
+      "loss": 0.4825,
+      "step": 306432
+    },
+    {
+      "epoch": 3.7678580752562345,
+      "grad_norm": 13.043724060058594,
+      "learning_rate": 1.623216128556686e-05,
+      "loss": 0.535,
+      "step": 306688
+    },
+    {
+      "epoch": 3.771003221932484,
+      "grad_norm": 5.998151779174805,
+      "learning_rate": 1.6229016167870658e-05,
+      "loss": 0.4982,
+      "step": 306944
+    },
+    {
+      "epoch": 3.774148368608733,
+      "grad_norm": 25.17442512512207,
+      "learning_rate": 1.6225871050174458e-05,
+      "loss": 0.5347,
+      "step": 307200
+    },
+    {
+      "epoch": 3.777293515284983,
+      "grad_norm": 83.40924835205078,
+      "learning_rate": 1.6222725932478254e-05,
+      "loss": 0.4761,
+      "step": 307456
+    },
+    {
+      "epoch": 3.7804386619612322,
+      "grad_norm": 10.433316230773926,
+      "learning_rate": 1.6219580814782054e-05,
+      "loss": 0.5412,
+      "step": 307712
+    },
+    {
+      "epoch": 3.783583808637482,
+      "grad_norm": 4.919463157653809,
+      "learning_rate": 1.6216435697085854e-05,
+      "loss": 0.5471,
+      "step": 307968
+    },
+    {
+      "epoch": 3.7867289553137313,
+      "grad_norm": 85.77107238769531,
+      "learning_rate": 1.621329057938965e-05,
+      "loss": 0.5045,
+      "step": 308224
+    },
+    {
+      "epoch": 3.789874101989981,
+      "grad_norm": 1.2499544620513916,
+      "learning_rate": 1.621014546169345e-05,
+      "loss": 0.4289,
+      "step": 308480
+    },
+    {
+      "epoch": 3.7930192486662304,
+      "grad_norm": 29.826250076293945,
+      "learning_rate": 1.620700034399725e-05,
+      "loss": 0.5874,
+      "step": 308736
+    },
+    {
+      "epoch": 3.79616439534248,
+      "grad_norm": 96.51860809326172,
+      "learning_rate": 1.620385522630105e-05,
+      "loss": 0.4276,
+      "step": 308992
+    },
+    {
+      "epoch": 3.7993095420187295,
+      "grad_norm": 29.25742530822754,
+      "learning_rate": 1.6200710108604846e-05,
+      "loss": 0.5067,
+      "step": 309248
+    },
+    {
+      "epoch": 3.8024546886949793,
+      "grad_norm": 1159.6026611328125,
+      "learning_rate": 1.6197564990908646e-05,
+      "loss": 0.5624,
+      "step": 309504
+    },
+    {
+      "epoch": 3.8055998353712286,
+      "grad_norm": 0.9833910465240479,
+      "learning_rate": 1.6194419873212443e-05,
+      "loss": 0.5727,
+      "step": 309760
+    },
+    {
+      "epoch": 3.8087449820474784,
+      "grad_norm": 102.49423217773438,
+      "learning_rate": 1.6191274755516242e-05,
+      "loss": 0.5264,
+      "step": 310016
+    },
+    {
+      "epoch": 3.8118901287237277,
+      "grad_norm": 37.246002197265625,
+      "learning_rate": 1.6188129637820042e-05,
+      "loss": 0.5061,
+      "step": 310272
+    },
+    {
+      "epoch": 3.8150352753999774,
+      "grad_norm": 108.58729553222656,
+      "learning_rate": 1.6184984520123842e-05,
+      "loss": 0.5599,
+      "step": 310528
+    },
+    {
+      "epoch": 3.8181804220762268,
+      "grad_norm": 10.775501251220703,
+      "learning_rate": 1.618183940242764e-05,
+      "loss": 0.6204,
+      "step": 310784
+    },
+    {
+      "epoch": 3.821325568752476,
+      "grad_norm": 15.554803848266602,
+      "learning_rate": 1.6178694284731438e-05,
+      "loss": 0.5168,
+      "step": 311040
+    },
+    {
+      "epoch": 3.824470715428726,
+      "grad_norm": 0.33082452416419983,
+      "learning_rate": 1.6175549167035235e-05,
+      "loss": 0.5284,
+      "step": 311296
+    },
+    {
+      "epoch": 3.8276158621049756,
+      "grad_norm": 0.38969486951828003,
+      "learning_rate": 1.6172404049339034e-05,
+      "loss": 0.5621,
+      "step": 311552
+    },
+    {
+      "epoch": 3.830761008781225,
+      "grad_norm": 143.03463745117188,
+      "learning_rate": 1.6169258931642834e-05,
+      "loss": 0.4814,
+      "step": 311808
+    },
+    {
+      "epoch": 3.8339061554574743,
+      "grad_norm": 0.08444385975599289,
+      "learning_rate": 1.6166113813946634e-05,
+      "loss": 0.4786,
+      "step": 312064
+    },
+    {
+      "epoch": 3.837051302133724,
+      "grad_norm": 4.929318428039551,
+      "learning_rate": 1.616296869625043e-05,
+      "loss": 0.4391,
+      "step": 312320
+    },
+    {
+      "epoch": 3.840196448809974,
+      "grad_norm": 144.34617614746094,
+      "learning_rate": 1.615982357855423e-05,
+      "loss": 0.6069,
+      "step": 312576
+    },
+    {
+      "epoch": 3.843341595486223,
+      "grad_norm": 0.1310855597257614,
+      "learning_rate": 1.615667846085803e-05,
+      "loss": 0.5042,
+      "step": 312832
+    },
+    {
+      "epoch": 3.8464867421624724,
+      "grad_norm": 0.05499180406332016,
+      "learning_rate": 1.6153533343161826e-05,
+      "loss": 0.4515,
+      "step": 313088
+    },
+    {
+      "epoch": 3.849631888838722,
+      "grad_norm": 16.251216888427734,
+      "learning_rate": 1.6150388225465626e-05,
+      "loss": 0.5915,
+      "step": 313344
+    },
+    {
+      "epoch": 3.852777035514972,
+      "grad_norm": 3.6784415245056152,
+      "learning_rate": 1.6147243107769426e-05,
+      "loss": 0.4512,
+      "step": 313600
+    },
+    {
+      "epoch": 3.8559221821912213,
+      "grad_norm": 19.981985092163086,
+      "learning_rate": 1.6144097990073226e-05,
+      "loss": 0.4899,
+      "step": 313856
+    },
+    {
+      "epoch": 3.8590673288674706,
+      "grad_norm": 12.532695770263672,
+      "learning_rate": 1.6140952872377022e-05,
+      "loss": 0.5655,
+      "step": 314112
+    },
+    {
+      "epoch": 3.8622124755437204,
+      "grad_norm": 12.418534278869629,
+      "learning_rate": 1.6137807754680822e-05,
+      "loss": 0.4427,
+      "step": 314368
+    },
+    {
+      "epoch": 3.8653576222199697,
+      "grad_norm": 286.6388854980469,
+      "learning_rate": 1.613466263698462e-05,
+      "loss": 0.5719,
+      "step": 314624
+    },
+    {
+      "epoch": 3.8685027688962195,
+      "grad_norm": 0.4477044939994812,
+      "learning_rate": 1.6131517519288418e-05,
+      "loss": 0.4793,
+      "step": 314880
+    },
+    {
+      "epoch": 3.871647915572469,
+      "grad_norm": 91.88975524902344,
+      "learning_rate": 1.6128372401592215e-05,
+      "loss": 0.5404,
+      "step": 315136
+    },
+    {
+      "epoch": 3.8747930622487186,
+      "grad_norm": 102.9833984375,
+      "learning_rate": 1.6125227283896018e-05,
+      "loss": 0.4834,
+      "step": 315392
+    },
+    {
+      "epoch": 3.877938208924968,
+      "grad_norm": 92.03607177734375,
+      "learning_rate": 1.6122082166199814e-05,
+      "loss": 0.5378,
+      "step": 315648
+    },
+    {
+      "epoch": 3.8810833556012176,
+      "grad_norm": 25.617658615112305,
+      "learning_rate": 1.6118937048503614e-05,
+      "loss": 0.4866,
+      "step": 315904
+    },
+    {
+      "epoch": 3.884228502277467,
+      "grad_norm": 17.501075744628906,
+      "learning_rate": 1.611579193080741e-05,
+      "loss": 0.5176,
+      "step": 316160
+    },
+    {
+      "epoch": 3.8873736489537167,
+      "grad_norm": 81.6054458618164,
+      "learning_rate": 1.611264681311121e-05,
+      "loss": 0.5075,
+      "step": 316416
+    },
+    {
+      "epoch": 3.890518795629966,
+      "grad_norm": 32.4160041809082,
+      "learning_rate": 1.610950169541501e-05,
+      "loss": 0.5325,
+      "step": 316672
+    },
+    {
+      "epoch": 3.893663942306216,
+      "grad_norm": 0.34635621309280396,
+      "learning_rate": 1.610635657771881e-05,
+      "loss": 0.5295,
+      "step": 316928
+    },
+    {
+      "epoch": 3.896809088982465,
+      "grad_norm": 2.071220636367798,
+      "learning_rate": 1.6103211460022606e-05,
+      "loss": 0.5175,
+      "step": 317184
+    },
+    {
+      "epoch": 3.899954235658715,
+      "grad_norm": 73.4476089477539,
+      "learning_rate": 1.6100066342326406e-05,
+      "loss": 0.4576,
+      "step": 317440
+    },
+    {
+      "epoch": 3.9030993823349642,
+      "grad_norm": 0.6312928199768066,
+      "learning_rate": 1.6096921224630203e-05,
+      "loss": 0.4924,
+      "step": 317696
+    },
+    {
+      "epoch": 3.9062445290112136,
+      "grad_norm": 47.5620002746582,
+      "learning_rate": 1.6093776106934002e-05,
+      "loss": 0.5348,
+      "step": 317952
+    },
+    {
+      "epoch": 3.9093896756874633,
+      "grad_norm": 0.6797400712966919,
+      "learning_rate": 1.6090630989237802e-05,
+      "loss": 0.5456,
+      "step": 318208
+    },
+    {
+      "epoch": 3.912534822363713,
+      "grad_norm": 9.67542839050293,
+      "learning_rate": 1.60874858715416e-05,
+      "loss": 0.6042,
+      "step": 318464
+    },
+    {
+      "epoch": 3.9156799690399624,
+      "grad_norm": 0.14112113416194916,
+      "learning_rate": 1.60843407538454e-05,
+      "loss": 0.4868,
+      "step": 318720
+    },
+    {
+      "epoch": 3.9188251157162117,
+      "grad_norm": 0.5000983476638794,
+      "learning_rate": 1.6081195636149198e-05,
+      "loss": 0.5176,
+      "step": 318976
+    },
+    {
+      "epoch": 3.9219702623924615,
+      "grad_norm": 0.8244889974594116,
+      "learning_rate": 1.6078050518452998e-05,
+      "loss": 0.5009,
+      "step": 319232
+    },
+    {
+      "epoch": 3.9251154090687113,
+      "grad_norm": 27.424039840698242,
+      "learning_rate": 1.6074905400756794e-05,
+      "loss": 0.5574,
+      "step": 319488
+    },
+    {
+      "epoch": 3.9282605557449606,
+      "grad_norm": 10.383527755737305,
+      "learning_rate": 1.6071760283060594e-05,
+      "loss": 0.5372,
+      "step": 319744
+    },
+    {
+      "epoch": 3.93140570242121,
+      "grad_norm": 68.16046142578125,
+      "learning_rate": 1.606861516536439e-05,
+      "loss": 0.5155,
+      "step": 320000
+    },
+    {
+      "epoch": 3.9345508490974597,
+      "grad_norm": 33.80850601196289,
+      "learning_rate": 1.6065470047668194e-05,
+      "loss": 0.5011,
+      "step": 320256
+    },
+    {
+      "epoch": 3.937695995773709,
+      "grad_norm": 116.3906478881836,
+      "learning_rate": 1.606232492997199e-05,
+      "loss": 0.4833,
+      "step": 320512
+    },
+    {
+      "epoch": 3.9408411424499588,
+      "grad_norm": 20.284738540649414,
+      "learning_rate": 1.605917981227579e-05,
+      "loss": 0.4831,
+      "step": 320768
+    },
+    {
+      "epoch": 3.943986289126208,
+      "grad_norm": 2.3668487071990967,
+      "learning_rate": 1.6056034694579586e-05,
+      "loss": 0.4916,
+      "step": 321024
+    },
+    {
+      "epoch": 3.947131435802458,
+      "grad_norm": 30.029603958129883,
+      "learning_rate": 1.6052889576883386e-05,
+      "loss": 0.555,
+      "step": 321280
+    },
+    {
+      "epoch": 3.950276582478707,
+      "grad_norm": 37.12919235229492,
+      "learning_rate": 1.6049744459187183e-05,
+      "loss": 0.5132,
+      "step": 321536
+    },
+    {
+      "epoch": 3.953421729154957,
+      "grad_norm": 2.277081251144409,
+      "learning_rate": 1.6046599341490983e-05,
+      "loss": 0.552,
+      "step": 321792
+    },
+    {
+      "epoch": 3.9565668758312063,
+      "grad_norm": 60.832698822021484,
+      "learning_rate": 1.6043454223794782e-05,
+      "loss": 0.474,
+      "step": 322048
+    },
+    {
+      "epoch": 3.959712022507456,
+      "grad_norm": 0.2835811972618103,
+      "learning_rate": 1.6040309106098582e-05,
+      "loss": 0.5508,
+      "step": 322304
+    },
+    {
+      "epoch": 3.9628571691837053,
+      "grad_norm": 0.07895316183567047,
+      "learning_rate": 1.603716398840238e-05,
+      "loss": 0.4787,
+      "step": 322560
+    },
+    {
+      "epoch": 3.966002315859955,
+      "grad_norm": 1.176039457321167,
+      "learning_rate": 1.603401887070618e-05,
+      "loss": 0.4864,
+      "step": 322816
+    },
+    {
+      "epoch": 3.9691474625362044,
+      "grad_norm": 47.2106819152832,
+      "learning_rate": 1.6030873753009978e-05,
+      "loss": 0.514,
+      "step": 323072
+    },
+    {
+      "epoch": 3.972292609212454,
+      "grad_norm": 3.9845356941223145,
+      "learning_rate": 1.6027728635313775e-05,
+      "loss": 0.6012,
+      "step": 323328
+    },
+    {
+      "epoch": 3.9754377558887035,
+      "grad_norm": 3.8236007690429688,
+      "learning_rate": 1.6024583517617574e-05,
+      "loss": 0.4826,
+      "step": 323584
+    },
+    {
+      "epoch": 3.978582902564953,
+      "grad_norm": 2.924356698989868,
+      "learning_rate": 1.6021438399921374e-05,
+      "loss": 0.4993,
+      "step": 323840
+    },
+    {
+      "epoch": 3.9817280492412026,
+      "grad_norm": 67.31713104248047,
+      "learning_rate": 1.6018293282225174e-05,
+      "loss": 0.4446,
+      "step": 324096
+    },
+    {
+      "epoch": 3.9848731959174524,
+      "grad_norm": 9.726791381835938,
+      "learning_rate": 1.601514816452897e-05,
+      "loss": 0.5299,
+      "step": 324352
+    },
+    {
+      "epoch": 3.9880183425937017,
+      "grad_norm": 0.436148464679718,
+      "learning_rate": 1.601200304683277e-05,
+      "loss": 0.4744,
+      "step": 324608
+    },
+    {
+      "epoch": 3.991163489269951,
+      "grad_norm": 18.131065368652344,
+      "learning_rate": 1.6008857929136567e-05,
+      "loss": 0.5299,
+      "step": 324864
+    },
+    {
+      "epoch": 3.994308635946201,
+      "grad_norm": 59.41075134277344,
+      "learning_rate": 1.6005712811440366e-05,
+      "loss": 0.5328,
+      "step": 325120
+    },
+    {
+      "epoch": 3.9974537826224505,
+      "grad_norm": 8.790979385375977,
+      "learning_rate": 1.6002567693744166e-05,
+      "loss": 0.5184,
+      "step": 325376
+    },
+    {
+      "epoch": 4.0,
+      "eval_AGE": {
+        "f1": 0.9615721314939476,
+        "number": 32192,
+        "precision": 0.9351770301215431,
+        "recall": 0.9895004970178927
+      },
+      "eval_BUILDINGNUM": {
+        "f1": 0.974103857162328,
+        "number": 28889,
+        "precision": 0.959103566276378,
+        "recall": 0.9895808093045796
+      },
+      "eval_CITY": {
+        "f1": 0.9711385680745608,
+        "number": 41669,
+        "precision": 0.9598031173092699,
+        "recall": 0.9827449662818882
+      },
+      "eval_DATE": {
+        "f1": 0.997630039252475,
+        "number": 80940,
+        "precision": 0.9967196132596685,
+        "recall": 0.9985421299728193
+      },
+      "eval_EMAIL": {
+        "f1": 0.9968215705523857,
+        "number": 51445,
+        "precision": 0.9969087798429116,
+        "recall": 0.9967343765186121
+      },
+      "eval_GENDER": {
+        "f1": 0.9717822063248737,
+        "number": 17457,
+        "precision": 0.963250604986212,
+        "recall": 0.980466288594833
+      },
+      "eval_GIVENNAME": {
+        "f1": 0.9305579573021605,
+        "number": 100350,
+        "precision": 0.922811590510441,
+        "recall": 0.938435475834579
+      },
+      "eval_IDCARDNUM": {
+        "f1": 0.9047753848833918,
+        "number": 22068,
+        "precision": 0.8708866065814295,
+        "recall": 0.9414083741163676
+      },
+      "eval_PASSPORTNUM": {
+        "f1": 0.9192574877576586,
+        "number": 12959,
+        "precision": 0.9046622833233712,
+        "recall": 0.934331352727834
+      },
+      "eval_SEX": {
+        "f1": 0.9667696532784071,
+        "number": 14570,
+        "precision": 0.9671016483516484,
+        "recall": 0.9664378860672614
+      },
+      "eval_SOCIALNUM": {
+        "f1": 0.9215204417068266,
+        "number": 15327,
+        "precision": 0.9019742596526302,
+        "recall": 0.9419325373523847
+      },
+      "eval_STREET": {
+        "f1": 0.9780866497361578,
+        "number": 32803,
+        "precision": 0.9758162398349314,
+        "recall": 0.9803676493003689
+      },
+      "eval_SURNAME": {
+        "f1": 0.9234893647042223,
+        "number": 85192,
+        "precision": 0.925066547313373,
+        "recall": 0.9219175509437506
+      },
+      "eval_TELEPHONENUM": {
+        "f1": 0.9934873665720848,
+        "number": 36941,
+        "precision": 0.9917721006771156,
+        "recall": 0.9952085758371457
+      },
+      "eval_ZIPCODE": {
+        "f1": 0.9801619764397906,
+        "number": 24346,
+        "precision": 0.9760896130346233,
+        "recall": 0.9842684629918672
+      },
+      "eval_overall_accuracy": 0.9928846432377629,
+      "eval_overall_f1": 0.9600317791152043,
+      "eval_overall_precision": 0.9528967694891431,
+      "eval_overall_recall": 0.9672744445263151,
+      "eval_runtime": 3733.3294,
+      "eval_samples_per_second": 87.21,
+      "eval_steps_per_second": 43.605,
+      "step": 325584
+    },
+    {
+      "epoch": 4.000589715001797,
+      "grad_norm": 35.143619537353516,
+      "learning_rate": 1.5999422576047966e-05,
+      "loss": 0.4844,
+      "step": 325632
+    },
+    {
+      "epoch": 4.003734861678046,
+      "grad_norm": 0.27741363644599915,
+      "learning_rate": 1.5996277458351762e-05,
+      "loss": 0.3744,
+      "step": 325888
+    },
+    {
+      "epoch": 4.006880008354296,
+      "grad_norm": 0.9082599878311157,
+      "learning_rate": 1.5993132340655562e-05,
+      "loss": 0.4088,
+      "step": 326144
+    },
+    {
+      "epoch": 4.0100251550305455,
+      "grad_norm": 6.128551959991455,
+      "learning_rate": 1.598998722295936e-05,
+      "loss": 0.3884,
+      "step": 326400
+    },
+    {
+      "epoch": 4.013170301706795,
+      "grad_norm": 119.51827239990234,
+      "learning_rate": 1.598684210526316e-05,
+      "loss": 0.4226,
+      "step": 326656
+    },
+    {
+      "epoch": 4.016315448383044,
+      "grad_norm": 175.86962890625,
+      "learning_rate": 1.5983696987566958e-05,
+      "loss": 0.4226,
+      "step": 326912
+    },
+    {
+      "epoch": 4.019460595059294,
+      "grad_norm": 96.47909545898438,
+      "learning_rate": 1.5980551869870758e-05,
+      "loss": 0.3977,
+      "step": 327168
+    },
+    {
+      "epoch": 4.022605741735544,
+      "grad_norm": 75.1969223022461,
+      "learning_rate": 1.5977406752174554e-05,
+      "loss": 0.3634,
+      "step": 327424
+    },
+    {
+      "epoch": 4.025750888411793,
+      "grad_norm": 0.052314598113298416,
+      "learning_rate": 1.5974261634478354e-05,
+      "loss": 0.4057,
+      "step": 327680
+    },
+    {
+      "epoch": 4.028896035088042,
+      "grad_norm": 9.135502815246582,
+      "learning_rate": 1.5971116516782154e-05,
+      "loss": 0.4358,
+      "step": 327936
+    },
+    {
+      "epoch": 4.032041181764292,
+      "grad_norm": 0.7194769978523254,
+      "learning_rate": 1.596797139908595e-05,
+      "loss": 0.3843,
+      "step": 328192
+    },
+    {
+      "epoch": 4.035186328440542,
+      "grad_norm": 1.8748400211334229,
+      "learning_rate": 1.596482628138975e-05,
+      "loss": 0.3709,
+      "step": 328448
+    },
+    {
+      "epoch": 4.038331475116792,
+      "grad_norm": 1.5528488159179688,
+      "learning_rate": 1.596168116369355e-05,
+      "loss": 0.4285,
+      "step": 328704
+    },
+    {
+      "epoch": 4.0414766217930405,
+      "grad_norm": 1.3449469804763794,
+      "learning_rate": 1.5958536045997347e-05,
+      "loss": 0.3928,
+      "step": 328960
+    },
+    {
+      "epoch": 4.04462176846929,
+      "grad_norm": 126.92698669433594,
+      "learning_rate": 1.5955390928301146e-05,
+      "loss": 0.4954,
+      "step": 329216
+    },
+    {
+      "epoch": 4.04776691514554,
+      "grad_norm": 2.424016237258911,
+      "learning_rate": 1.5952245810604946e-05,
+      "loss": 0.4058,
+      "step": 329472
+    },
+    {
+      "epoch": 4.05091206182179,
+      "grad_norm": 0.7729480862617493,
+      "learning_rate": 1.5949100692908743e-05,
+      "loss": 0.5069,
+      "step": 329728
+    },
+    {
+      "epoch": 4.054057208498039,
+      "grad_norm": 7.9247589111328125,
+      "learning_rate": 1.5945955575212542e-05,
+      "loss": 0.4159,
+      "step": 329984
+    },
+    {
+      "epoch": 4.057202355174288,
+      "grad_norm": 799.33837890625,
+      "learning_rate": 1.5942810457516342e-05,
+      "loss": 0.4731,
+      "step": 330240
+    },
+    {
+      "epoch": 4.060347501850538,
+      "grad_norm": 1.0703004598617554,
+      "learning_rate": 1.5939665339820142e-05,
+      "loss": 0.3643,
+      "step": 330496
+    },
+    {
+      "epoch": 4.063492648526788,
+      "grad_norm": 16.651174545288086,
+      "learning_rate": 1.593652022212394e-05,
+      "loss": 0.408,
+      "step": 330752
+    },
+    {
+      "epoch": 4.066637795203037,
+      "grad_norm": 133.6138153076172,
+      "learning_rate": 1.5933375104427738e-05,
+      "loss": 0.4181,
+      "step": 331008
+    },
+    {
+      "epoch": 4.069782941879287,
+      "grad_norm": 0.37095826864242554,
+      "learning_rate": 1.5930229986731535e-05,
+      "loss": 0.4545,
+      "step": 331264
+    },
+    {
+      "epoch": 4.072928088555536,
+      "grad_norm": 59.24568557739258,
+      "learning_rate": 1.5927084869035334e-05,
+      "loss": 0.3876,
+      "step": 331520
+    },
+    {
+      "epoch": 4.076073235231785,
+      "grad_norm": 0.05594320222735405,
+      "learning_rate": 1.592393975133913e-05,
+      "loss": 0.4537,
+      "step": 331776
+    },
+    {
+      "epoch": 4.079218381908035,
+      "grad_norm": 1.0508644580841064,
+      "learning_rate": 1.5920794633642934e-05,
+      "loss": 0.4853,
+      "step": 332032
+    },
+    {
+      "epoch": 4.082363528584285,
+      "grad_norm": 49.64292526245117,
+      "learning_rate": 1.591764951594673e-05,
+      "loss": 0.4517,
+      "step": 332288
+    },
+    {
+      "epoch": 4.0855086752605345,
+      "grad_norm": 9.851983070373535,
+      "learning_rate": 1.591450439825053e-05,
+      "loss": 0.4809,
+      "step": 332544
+    },
+    {
+      "epoch": 4.088653821936783,
+      "grad_norm": 17.43916130065918,
+      "learning_rate": 1.5911359280554327e-05,
+      "loss": 0.4832,
+      "step": 332800
+    },
+    {
+      "epoch": 4.091798968613033,
+      "grad_norm": 107.35646057128906,
+      "learning_rate": 1.5908214162858126e-05,
+      "loss": 0.4221,
+      "step": 333056
+    },
+    {
+      "epoch": 4.094944115289283,
+      "grad_norm": 0.1817217767238617,
+      "learning_rate": 1.5905069045161926e-05,
+      "loss": 0.4883,
+      "step": 333312
+    },
+    {
+      "epoch": 4.098089261965533,
+      "grad_norm": 143.05455017089844,
+      "learning_rate": 1.5901923927465723e-05,
+      "loss": 0.4395,
+      "step": 333568
+    },
+    {
+      "epoch": 4.101234408641782,
+      "grad_norm": 35.979209899902344,
+      "learning_rate": 1.5898778809769522e-05,
+      "loss": 0.4266,
+      "step": 333824
+    },
+    {
+      "epoch": 4.104379555318031,
+      "grad_norm": 0.00816342793405056,
+      "learning_rate": 1.5895633692073322e-05,
+      "loss": 0.4351,
+      "step": 334080
+    },
+    {
+      "epoch": 4.107524701994281,
+      "grad_norm": 213.85360717773438,
+      "learning_rate": 1.5892488574377122e-05,
+      "loss": 0.3684,
+      "step": 334336
+    },
+    {
+      "epoch": 4.110669848670531,
+      "grad_norm": 39.7363395690918,
+      "learning_rate": 1.588934345668092e-05,
+      "loss": 0.3954,
+      "step": 334592
+    },
+    {
+      "epoch": 4.11381499534678,
+      "grad_norm": 0.14550000429153442,
+      "learning_rate": 1.588619833898472e-05,
+      "loss": 0.409,
+      "step": 334848
+    },
+    {
+      "epoch": 4.1169601420230295,
+      "grad_norm": 2.4026780128479004,
+      "learning_rate": 1.5883053221288515e-05,
+      "loss": 0.4363,
+      "step": 335104
+    },
+    {
+      "epoch": 4.120105288699279,
+      "grad_norm": 147.01051330566406,
+      "learning_rate": 1.5879908103592318e-05,
+      "loss": 0.5452,
+      "step": 335360
+    },
+    {
+      "epoch": 4.123250435375529,
+      "grad_norm": 0.21769696474075317,
+      "learning_rate": 1.5876762985896114e-05,
+      "loss": 0.3588,
+      "step": 335616
+    },
+    {
+      "epoch": 4.126395582051778,
+      "grad_norm": 0.3982313275337219,
+      "learning_rate": 1.5873617868199914e-05,
+      "loss": 0.4838,
+      "step": 335872
+    },
+    {
+      "epoch": 4.129540728728028,
+      "grad_norm": 38.10209655761719,
+      "learning_rate": 1.587047275050371e-05,
+      "loss": 0.3893,
+      "step": 336128
+    },
+    {
+      "epoch": 4.1326858754042775,
+      "grad_norm": 4.889694690704346,
+      "learning_rate": 1.586732763280751e-05,
+      "loss": 0.4769,
+      "step": 336384
+    },
+    {
+      "epoch": 4.135831022080527,
+      "grad_norm": 46.47163391113281,
+      "learning_rate": 1.5864182515111307e-05,
+      "loss": 0.472,
+      "step": 336640
+    },
+    {
+      "epoch": 4.138976168756776,
+      "grad_norm": 0.6901742219924927,
+      "learning_rate": 1.5861037397415107e-05,
+      "loss": 0.4908,
+      "step": 336896
+    },
+    {
+      "epoch": 4.142121315433026,
+      "grad_norm": 28.333667755126953,
+      "learning_rate": 1.5857892279718906e-05,
+      "loss": 0.3538,
+      "step": 337152
+    },
+    {
+      "epoch": 4.145266462109276,
+      "grad_norm": 0.4067509174346924,
+      "learning_rate": 1.5854747162022706e-05,
+      "loss": 0.4117,
+      "step": 337408
+    },
+    {
+      "epoch": 4.1484116087855245,
+      "grad_norm": 18.79134750366211,
+      "learning_rate": 1.5851602044326503e-05,
+      "loss": 0.4897,
+      "step": 337664
+    },
+    {
+      "epoch": 4.151556755461774,
+      "grad_norm": 24.368175506591797,
+      "learning_rate": 1.5848456926630302e-05,
+      "loss": 0.5052,
+      "step": 337920
+    },
+    {
+      "epoch": 4.154701902138024,
+      "grad_norm": 34.97966766357422,
+      "learning_rate": 1.5845311808934102e-05,
+      "loss": 0.3691,
+      "step": 338176
+    },
+    {
+      "epoch": 4.157847048814274,
+      "grad_norm": 36.493995666503906,
+      "learning_rate": 1.58421666912379e-05,
+      "loss": 0.4531,
+      "step": 338432
+    },
+    {
+      "epoch": 4.160992195490523,
+      "grad_norm": 143.27223205566406,
+      "learning_rate": 1.58390215735417e-05,
+      "loss": 0.4265,
+      "step": 338688
+    },
+    {
+      "epoch": 4.1641373421667724,
+      "grad_norm": 0.828838050365448,
+      "learning_rate": 1.5835876455845498e-05,
+      "loss": 0.3934,
+      "step": 338944
+    },
+    {
+      "epoch": 4.167282488843022,
+      "grad_norm": 0.25666344165802,
+      "learning_rate": 1.5832731338149298e-05,
+      "loss": 0.4192,
+      "step": 339200
+    },
+    {
+      "epoch": 4.170427635519272,
+      "grad_norm": 0.1353408545255661,
+      "learning_rate": 1.5829586220453094e-05,
+      "loss": 0.3556,
+      "step": 339456
+    },
+    {
+      "epoch": 4.173572782195521,
+      "grad_norm": 14.724331855773926,
+      "learning_rate": 1.5826441102756894e-05,
+      "loss": 0.4295,
+      "step": 339712
+    },
+    {
+      "epoch": 4.176717928871771,
+      "grad_norm": 0.2672252058982849,
+      "learning_rate": 1.582329598506069e-05,
+      "loss": 0.4385,
+      "step": 339968
+    },
+    {
+      "epoch": 4.17986307554802,
+      "grad_norm": 0.03450535982847214,
+      "learning_rate": 1.582015086736449e-05,
+      "loss": 0.4235,
+      "step": 340224
+    },
+    {
+      "epoch": 4.18300822222427,
+      "grad_norm": 0.24874207377433777,
+      "learning_rate": 1.581700574966829e-05,
+      "loss": 0.4061,
+      "step": 340480
+    },
+    {
+      "epoch": 4.186153368900519,
+      "grad_norm": 0.7424083948135376,
+      "learning_rate": 1.581386063197209e-05,
+      "loss": 0.4908,
+      "step": 340736
+    },
+    {
+      "epoch": 4.189298515576769,
+      "grad_norm": 134.73683166503906,
+      "learning_rate": 1.5810715514275887e-05,
+      "loss": 0.429,
+      "step": 340992
+    },
+    {
+      "epoch": 4.192443662253019,
+      "grad_norm": 5.836315631866455,
+      "learning_rate": 1.5807570396579686e-05,
+      "loss": 0.4462,
+      "step": 341248
+    },
+    {
+      "epoch": 4.195588808929268,
+      "grad_norm": 0.018218902871012688,
+      "learning_rate": 1.5804425278883483e-05,
+      "loss": 0.445,
+      "step": 341504
+    },
+    {
+      "epoch": 4.198733955605517,
+      "grad_norm": 0.7838549017906189,
+      "learning_rate": 1.5801280161187283e-05,
+      "loss": 0.4811,
+      "step": 341760
+    },
+    {
+      "epoch": 4.201879102281767,
+      "grad_norm": 1.8249409198760986,
+      "learning_rate": 1.5798135043491082e-05,
+      "loss": 0.347,
+      "step": 342016
+    },
+    {
+      "epoch": 4.205024248958017,
+      "grad_norm": 566.2861328125,
+      "learning_rate": 1.5794989925794882e-05,
+      "loss": 0.4983,
+      "step": 342272
+    },
+    {
+      "epoch": 4.2081693956342665,
+      "grad_norm": 0.07192770391702652,
+      "learning_rate": 1.579184480809868e-05,
+      "loss": 0.4245,
+      "step": 342528
+    },
+    {
+      "epoch": 4.211314542310515,
+      "grad_norm": 0.33272820711135864,
+      "learning_rate": 1.578869969040248e-05,
+      "loss": 0.4771,
+      "step": 342784
+    },
+    {
+      "epoch": 4.214459688986765,
+      "grad_norm": 13.047657012939453,
+      "learning_rate": 1.5785554572706275e-05,
+      "loss": 0.4365,
+      "step": 343040
+    },
+    {
+      "epoch": 4.217604835663015,
+      "grad_norm": 229.12564086914062,
+      "learning_rate": 1.5782409455010075e-05,
+      "loss": 0.3834,
+      "step": 343296
+    },
+    {
+      "epoch": 4.220749982339265,
+      "grad_norm": 48.16639709472656,
+      "learning_rate": 1.5779264337313874e-05,
+      "loss": 0.4109,
+      "step": 343552
+    },
+    {
+      "epoch": 4.223895129015514,
+      "grad_norm": 0.33954524993896484,
+      "learning_rate": 1.5776119219617674e-05,
+      "loss": 0.4298,
+      "step": 343808
+    },
+    {
+      "epoch": 4.227040275691763,
+      "grad_norm": 18.813026428222656,
+      "learning_rate": 1.577297410192147e-05,
+      "loss": 0.4408,
+      "step": 344064
+    },
+    {
+      "epoch": 4.230185422368013,
+      "grad_norm": 0.6447019577026367,
+      "learning_rate": 1.576982898422527e-05,
+      "loss": 0.4882,
+      "step": 344320
+    },
+    {
+      "epoch": 4.233330569044263,
+      "grad_norm": 0.34302687644958496,
+      "learning_rate": 1.576668386652907e-05,
+      "loss": 0.3933,
+      "step": 344576
+    },
+    {
+      "epoch": 4.236475715720512,
+      "grad_norm": 16.732297897338867,
+      "learning_rate": 1.5763538748832867e-05,
+      "loss": 0.413,
+      "step": 344832
+    },
+    {
+      "epoch": 4.2396208623967615,
+      "grad_norm": 55.835086822509766,
+      "learning_rate": 1.5760393631136666e-05,
+      "loss": 0.4081,
+      "step": 345088
+    },
+    {
+      "epoch": 4.242766009073011,
+      "grad_norm": 0.14344307780265808,
+      "learning_rate": 1.5757248513440466e-05,
+      "loss": 0.4701,
+      "step": 345344
+    },
+    {
+      "epoch": 4.24591115574926,
+      "grad_norm": 19.056779861450195,
+      "learning_rate": 1.5754103395744266e-05,
+      "loss": 0.4397,
+      "step": 345600
+    },
+    {
+      "epoch": 4.24905630242551,
+      "grad_norm": 0.48494040966033936,
+      "learning_rate": 1.5750958278048062e-05,
+      "loss": 0.4442,
+      "step": 345856
+    },
+    {
+      "epoch": 4.25220144910176,
+      "grad_norm": 121.74722290039062,
+      "learning_rate": 1.5747813160351862e-05,
+      "loss": 0.4539,
+      "step": 346112
+    },
+    {
+      "epoch": 4.2553465957780094,
+      "grad_norm": 73.08151245117188,
+      "learning_rate": 1.574466804265566e-05,
+      "loss": 0.3675,
+      "step": 346368
+    },
+    {
+      "epoch": 4.258491742454258,
+      "grad_norm": 0.5641027092933655,
+      "learning_rate": 1.574152292495946e-05,
+      "loss": 0.4251,
+      "step": 346624
+    },
+    {
+      "epoch": 4.261636889130508,
+      "grad_norm": 193.56671142578125,
+      "learning_rate": 1.5738377807263255e-05,
+      "loss": 0.3765,
+      "step": 346880
+    },
+    {
+      "epoch": 4.264782035806758,
+      "grad_norm": 0.43298056721687317,
+      "learning_rate": 1.5735232689567058e-05,
+      "loss": 0.416,
+      "step": 347136
+    },
+    {
+      "epoch": 4.267927182483008,
+      "grad_norm": 0.6539829969406128,
+      "learning_rate": 1.5732087571870855e-05,
+      "loss": 0.4428,
+      "step": 347392
+    },
+    {
+      "epoch": 4.2710723291592565,
+      "grad_norm": 88.20236206054688,
+      "learning_rate": 1.5728942454174654e-05,
+      "loss": 0.494,
+      "step": 347648
+    },
+    {
+      "epoch": 4.274217475835506,
+      "grad_norm": 0.46211355924606323,
+      "learning_rate": 1.572579733647845e-05,
+      "loss": 0.3858,
+      "step": 347904
+    },
+    {
+      "epoch": 4.277362622511756,
+      "grad_norm": 10.243553161621094,
+      "learning_rate": 1.572265221878225e-05,
+      "loss": 0.4337,
+      "step": 348160
+    },
+    {
+      "epoch": 4.280507769188006,
+      "grad_norm": 1.9267185926437378,
+      "learning_rate": 1.571950710108605e-05,
+      "loss": 0.5603,
+      "step": 348416
+    },
+    {
+      "epoch": 4.283652915864255,
+      "grad_norm": 31.92858123779297,
+      "learning_rate": 1.571636198338985e-05,
+      "loss": 0.4218,
+      "step": 348672
+    },
+    {
+      "epoch": 4.286798062540504,
+      "grad_norm": 36.23369598388672,
+      "learning_rate": 1.5713216865693647e-05,
+      "loss": 0.4761,
+      "step": 348928
+    },
+    {
+      "epoch": 4.289943209216754,
+      "grad_norm": 94.79277038574219,
+      "learning_rate": 1.5710071747997446e-05,
+      "loss": 0.5288,
+      "step": 349184
+    },
+    {
+      "epoch": 4.293088355893004,
+      "grad_norm": 1.9670157432556152,
+      "learning_rate": 1.5706926630301246e-05,
+      "loss": 0.4052,
+      "step": 349440
+    },
+    {
+      "epoch": 4.296233502569253,
+      "grad_norm": 122.4579849243164,
+      "learning_rate": 1.5703781512605043e-05,
+      "loss": 0.4793,
+      "step": 349696
+    },
+    {
+      "epoch": 4.299378649245503,
+      "grad_norm": 42.34500503540039,
+      "learning_rate": 1.5700636394908842e-05,
+      "loss": 0.3667,
+      "step": 349952
+    },
+    {
+      "epoch": 4.302523795921752,
+      "grad_norm": 338.1840515136719,
+      "learning_rate": 1.569749127721264e-05,
+      "loss": 0.5043,
+      "step": 350208
+    },
+    {
+      "epoch": 4.305668942598002,
+      "grad_norm": 0.07519801706075668,
+      "learning_rate": 1.5694346159516442e-05,
+      "loss": 0.3691,
+      "step": 350464
+    },
+    {
+      "epoch": 4.308814089274251,
+      "grad_norm": 0.30865201354026794,
+      "learning_rate": 1.569120104182024e-05,
+      "loss": 0.4798,
+      "step": 350720
+    },
+    {
+      "epoch": 4.311959235950501,
+      "grad_norm": 13.652377128601074,
+      "learning_rate": 1.5688055924124038e-05,
+      "loss": 0.4618,
+      "step": 350976
+    },
+    {
+      "epoch": 4.315104382626751,
+      "grad_norm": 80.21342468261719,
+      "learning_rate": 1.5684910806427835e-05,
+      "loss": 0.4714,
+      "step": 351232
+    },
+    {
+      "epoch": 4.318249529302999,
+      "grad_norm": 0.2811349332332611,
+      "learning_rate": 1.5681765688731634e-05,
+      "loss": 0.3776,
+      "step": 351488
+    },
+    {
+      "epoch": 4.321394675979249,
+      "grad_norm": 43.756290435791016,
+      "learning_rate": 1.567862057103543e-05,
+      "loss": 0.5404,
+      "step": 351744
+    },
+    {
+      "epoch": 4.324539822655499,
+      "grad_norm": 7.5840630531311035,
+      "learning_rate": 1.5675475453339234e-05,
+      "loss": 0.3703,
+      "step": 352000
+    },
+    {
+      "epoch": 4.327684969331749,
+      "grad_norm": 139.19482421875,
+      "learning_rate": 1.567233033564303e-05,
+      "loss": 0.3948,
+      "step": 352256
+    },
+    {
+      "epoch": 4.330830116007998,
+      "grad_norm": 104.86837768554688,
+      "learning_rate": 1.566918521794683e-05,
+      "loss": 0.4202,
+      "step": 352512
+    },
+    {
+      "epoch": 4.333975262684247,
+      "grad_norm": 3.997427463531494,
+      "learning_rate": 1.5666040100250627e-05,
+      "loss": 0.4101,
+      "step": 352768
+    },
+    {
+      "epoch": 4.337120409360497,
+      "grad_norm": 0.5823514461517334,
+      "learning_rate": 1.5662894982554426e-05,
+      "loss": 0.4543,
+      "step": 353024
+    },
+    {
+      "epoch": 4.340265556036747,
+      "grad_norm": 0.5827736258506775,
+      "learning_rate": 1.5659749864858226e-05,
+      "loss": 0.4708,
+      "step": 353280
+    },
+    {
+      "epoch": 4.343410702712996,
+      "grad_norm": 0.031142117455601692,
+      "learning_rate": 1.5656604747162023e-05,
+      "loss": 0.4162,
+      "step": 353536
+    },
+    {
+      "epoch": 4.3465558493892456,
+      "grad_norm": 0.04433192312717438,
+      "learning_rate": 1.5653459629465823e-05,
+      "loss": 0.4091,
+      "step": 353792
+    },
+    {
+      "epoch": 4.349700996065495,
+      "grad_norm": 17.527206420898438,
+      "learning_rate": 1.5650314511769622e-05,
+      "loss": 0.4452,
+      "step": 354048
+    },
+    {
+      "epoch": 4.352846142741745,
+      "grad_norm": 26.186925888061523,
+      "learning_rate": 1.564716939407342e-05,
+      "loss": 0.4063,
+      "step": 354304
+    },
+    {
+      "epoch": 4.355991289417994,
+      "grad_norm": 0.2337101846933365,
+      "learning_rate": 1.564402427637722e-05,
+      "loss": 0.403,
+      "step": 354560
+    },
+    {
+      "epoch": 4.359136436094244,
+      "grad_norm": 51.35133743286133,
+      "learning_rate": 1.564087915868102e-05,
+      "loss": 0.5557,
+      "step": 354816
+    },
+    {
+      "epoch": 4.3622815827704935,
+      "grad_norm": 2.160275936126709,
+      "learning_rate": 1.5637734040984815e-05,
+      "loss": 0.394,
+      "step": 355072
+    },
+    {
+      "epoch": 4.365426729446743,
+      "grad_norm": 0.6121172308921814,
+      "learning_rate": 1.5634588923288615e-05,
+      "loss": 0.445,
+      "step": 355328
+    },
+    {
+      "epoch": 4.368571876122992,
+      "grad_norm": 14.309392929077148,
+      "learning_rate": 1.5631443805592414e-05,
+      "loss": 0.4381,
+      "step": 355584
+    },
+    {
+      "epoch": 4.371717022799242,
+      "grad_norm": 0.30734115839004517,
+      "learning_rate": 1.5628298687896214e-05,
+      "loss": 0.463,
+      "step": 355840
+    },
+    {
+      "epoch": 4.374862169475492,
+      "grad_norm": 0.7233496904373169,
+      "learning_rate": 1.562515357020001e-05,
+      "loss": 0.5224,
+      "step": 356096
+    },
+    {
+      "epoch": 4.378007316151741,
+      "grad_norm": 10.601611137390137,
+      "learning_rate": 1.562200845250381e-05,
+      "loss": 0.3838,
+      "step": 356352
+    },
+    {
+      "epoch": 4.38115246282799,
+      "grad_norm": 209.69227600097656,
+      "learning_rate": 1.5618863334807607e-05,
+      "loss": 0.4121,
+      "step": 356608
+    },
+    {
+      "epoch": 4.38429760950424,
+      "grad_norm": 75.78793334960938,
+      "learning_rate": 1.5615718217111407e-05,
+      "loss": 0.4433,
+      "step": 356864
+    },
+    {
+      "epoch": 4.38744275618049,
+      "grad_norm": 3.6077990531921387,
+      "learning_rate": 1.5612573099415206e-05,
+      "loss": 0.485,
+      "step": 357120
+    },
+    {
+      "epoch": 4.390587902856739,
+      "grad_norm": 6.1350812911987305,
+      "learning_rate": 1.5609427981719006e-05,
+      "loss": 0.5076,
+      "step": 357376
+    },
+    {
+      "epoch": 4.3937330495329885,
+      "grad_norm": 12.99637508392334,
+      "learning_rate": 1.5606282864022803e-05,
+      "loss": 0.4506,
+      "step": 357632
+    },
+    {
+      "epoch": 4.396878196209238,
+      "grad_norm": 0.5122680068016052,
+      "learning_rate": 1.5603137746326602e-05,
+      "loss": 0.4193,
+      "step": 357888
+    },
+    {
+      "epoch": 4.400023342885488,
+      "grad_norm": 0.025185653939843178,
+      "learning_rate": 1.55999926286304e-05,
+      "loss": 0.3413,
+      "step": 358144
+    },
+    {
+      "epoch": 4.403168489561738,
+      "grad_norm": 3.367849111557007,
+      "learning_rate": 1.55968475109342e-05,
+      "loss": 0.4435,
+      "step": 358400
+    },
+    {
+      "epoch": 4.406313636237987,
+      "grad_norm": 27.118183135986328,
+      "learning_rate": 1.5593702393238e-05,
+      "loss": 0.464,
+      "step": 358656
+    },
+    {
+      "epoch": 4.409458782914236,
+      "grad_norm": 4.735858917236328,
+      "learning_rate": 1.5590557275541798e-05,
+      "loss": 0.4544,
+      "step": 358912
+    },
+    {
+      "epoch": 4.412603929590486,
+      "grad_norm": 13.226275444030762,
+      "learning_rate": 1.5587412157845595e-05,
+      "loss": 0.3296,
+      "step": 359168
+    },
+    {
+      "epoch": 4.415749076266735,
+      "grad_norm": 0.22179430723190308,
+      "learning_rate": 1.5584267040149394e-05,
+      "loss": 0.5463,
+      "step": 359424
+    },
+    {
+      "epoch": 4.418894222942985,
+      "grad_norm": 6.1947712898254395,
+      "learning_rate": 1.5581121922453194e-05,
+      "loss": 0.3669,
+      "step": 359680
+    },
+    {
+      "epoch": 4.422039369619235,
+      "grad_norm": 25.431640625,
+      "learning_rate": 1.557797680475699e-05,
+      "loss": 0.4295,
+      "step": 359936
+    },
+    {
+      "epoch": 4.425184516295484,
+      "grad_norm": 4.872597694396973,
+      "learning_rate": 1.557483168706079e-05,
+      "loss": 0.4799,
+      "step": 360192
+    },
+    {
+      "epoch": 4.428329662971733,
+      "grad_norm": 337.6877136230469,
+      "learning_rate": 1.557168656936459e-05,
+      "loss": 0.3237,
+      "step": 360448
+    },
+    {
+      "epoch": 4.431474809647983,
+      "grad_norm": 3.2726728916168213,
+      "learning_rate": 1.556854145166839e-05,
+      "loss": 0.3633,
+      "step": 360704
+    },
+    {
+      "epoch": 4.434619956324233,
+      "grad_norm": 0.7783453464508057,
+      "learning_rate": 1.5565396333972187e-05,
+      "loss": 0.3915,
+      "step": 360960
+    },
+    {
+      "epoch": 4.4377651030004825,
+      "grad_norm": 0.5727116465568542,
+      "learning_rate": 1.5562251216275986e-05,
+      "loss": 0.5754,
+      "step": 361216
+    },
+    {
+      "epoch": 4.440910249676731,
+      "grad_norm": 0.19325481355190277,
+      "learning_rate": 1.5559106098579783e-05,
+      "loss": 0.4565,
+      "step": 361472
+    },
+    {
+      "epoch": 4.444055396352981,
+      "grad_norm": 0.3410254120826721,
+      "learning_rate": 1.5555960980883583e-05,
+      "loss": 0.4259,
+      "step": 361728
+    },
+    {
+      "epoch": 4.447200543029231,
+      "grad_norm": 5.404856204986572,
+      "learning_rate": 1.555281586318738e-05,
+      "loss": 0.4012,
+      "step": 361984
+    },
+    {
+      "epoch": 4.450345689705481,
+      "grad_norm": 0.1795419603586197,
+      "learning_rate": 1.5549670745491182e-05,
+      "loss": 0.4952,
+      "step": 362240
+    },
+    {
+      "epoch": 4.45349083638173,
+      "grad_norm": 0.17945712804794312,
+      "learning_rate": 1.554652562779498e-05,
+      "loss": 0.371,
+      "step": 362496
+    },
+    {
+      "epoch": 4.456635983057979,
+      "grad_norm": 11.178935050964355,
+      "learning_rate": 1.554338051009878e-05,
+      "loss": 0.4235,
+      "step": 362752
+    },
+    {
+      "epoch": 4.459781129734229,
+      "grad_norm": 19.169958114624023,
+      "learning_rate": 1.5540235392402575e-05,
+      "loss": 0.4337,
+      "step": 363008
+    },
+    {
+      "epoch": 4.462926276410479,
+      "grad_norm": 40.609474182128906,
+      "learning_rate": 1.5537090274706375e-05,
+      "loss": 0.4874,
+      "step": 363264
+    },
+    {
+      "epoch": 4.466071423086728,
+      "grad_norm": 0.4823182225227356,
+      "learning_rate": 1.5533945157010174e-05,
+      "loss": 0.4639,
+      "step": 363520
+    },
+    {
+      "epoch": 4.4692165697629775,
+      "grad_norm": 0.10706117749214172,
+      "learning_rate": 1.5530800039313974e-05,
+      "loss": 0.4095,
+      "step": 363776
+    },
+    {
+      "epoch": 4.472361716439227,
+      "grad_norm": 0.14544102549552917,
+      "learning_rate": 1.552765492161777e-05,
+      "loss": 0.4331,
+      "step": 364032
+    },
+    {
+      "epoch": 4.475506863115477,
+      "grad_norm": 78.92945861816406,
+      "learning_rate": 1.552450980392157e-05,
+      "loss": 0.3828,
+      "step": 364288
+    },
+    {
+      "epoch": 4.478652009791726,
+      "grad_norm": 0.9381522536277771,
+      "learning_rate": 1.552136468622537e-05,
+      "loss": 0.5498,
+      "step": 364544
+    },
+    {
+      "epoch": 4.481797156467976,
+      "grad_norm": 48.698570251464844,
+      "learning_rate": 1.5518219568529167e-05,
+      "loss": 0.4764,
+      "step": 364800
+    },
+    {
+      "epoch": 4.4849423031442255,
+      "grad_norm": 101.31189727783203,
+      "learning_rate": 1.5515074450832966e-05,
+      "loss": 0.5105,
+      "step": 365056
+    },
+    {
+      "epoch": 4.488087449820474,
+      "grad_norm": 0.1653733104467392,
+      "learning_rate": 1.5511929333136763e-05,
+      "loss": 0.4883,
+      "step": 365312
+    },
+    {
+      "epoch": 4.491232596496724,
+      "grad_norm": 14.87435531616211,
+      "learning_rate": 1.5508784215440563e-05,
+      "loss": 0.4209,
+      "step": 365568
+    },
+    {
+      "epoch": 4.494377743172974,
+      "grad_norm": 64.11092376708984,
+      "learning_rate": 1.5505639097744362e-05,
+      "loss": 0.4409,
+      "step": 365824
+    },
+    {
+      "epoch": 4.497522889849224,
+      "grad_norm": 0.4967323839664459,
+      "learning_rate": 1.5502493980048162e-05,
+      "loss": 0.4011,
+      "step": 366080
+    },
+    {
+      "epoch": 4.5006680365254725,
+      "grad_norm": 32.29637145996094,
+      "learning_rate": 1.549934886235196e-05,
+      "loss": 0.5045,
+      "step": 366336
+    },
+    {
+      "epoch": 4.503813183201722,
+      "grad_norm": 17.688823699951172,
+      "learning_rate": 1.549620374465576e-05,
+      "loss": 0.4361,
+      "step": 366592
+    },
+    {
+      "epoch": 4.506958329877972,
+      "grad_norm": 2.3698978424072266,
+      "learning_rate": 1.5493058626959555e-05,
+      "loss": 0.4353,
+      "step": 366848
+    },
+    {
+      "epoch": 4.510103476554222,
+      "grad_norm": 0.2591179609298706,
+      "learning_rate": 1.5489913509263358e-05,
+      "loss": 0.5272,
+      "step": 367104
+    },
+    {
+      "epoch": 4.513248623230471,
+      "grad_norm": 105.75251770019531,
+      "learning_rate": 1.5486768391567155e-05,
+      "loss": 0.3966,
+      "step": 367360
+    },
+    {
+      "epoch": 4.5163937699067205,
+      "grad_norm": 42.855709075927734,
+      "learning_rate": 1.5483623273870954e-05,
+      "loss": 0.4623,
+      "step": 367616
+    },
+    {
+      "epoch": 4.51953891658297,
+      "grad_norm": 64.96548461914062,
+      "learning_rate": 1.548047815617475e-05,
+      "loss": 0.5504,
+      "step": 367872
+    },
+    {
+      "epoch": 4.52268406325922,
+      "grad_norm": 40.712425231933594,
+      "learning_rate": 1.547733303847855e-05,
+      "loss": 0.4,
+      "step": 368128
+    },
+    {
+      "epoch": 4.525829209935469,
+      "grad_norm": 0.04637348651885986,
+      "learning_rate": 1.5474187920782347e-05,
+      "loss": 0.5036,
+      "step": 368384
+    },
+    {
+      "epoch": 4.528974356611719,
+      "grad_norm": 42.304561614990234,
+      "learning_rate": 1.5471042803086147e-05,
+      "loss": 0.4588,
+      "step": 368640
+    },
+    {
+      "epoch": 4.532119503287968,
+      "grad_norm": 2.494689464569092,
+      "learning_rate": 1.5467897685389947e-05,
+      "loss": 0.4567,
+      "step": 368896
+    },
+    {
+      "epoch": 4.535264649964218,
+      "grad_norm": 0.09282395988702774,
+      "learning_rate": 1.5464752567693746e-05,
+      "loss": 0.4844,
+      "step": 369152
+    },
+    {
+      "epoch": 4.538409796640467,
+      "grad_norm": 91.8843765258789,
+      "learning_rate": 1.5461607449997543e-05,
+      "loss": 0.407,
+      "step": 369408
+    },
+    {
+      "epoch": 4.541554943316717,
+      "grad_norm": 1.5122617483139038,
+      "learning_rate": 1.5458462332301343e-05,
+      "loss": 0.4167,
+      "step": 369664
+    },
+    {
+      "epoch": 4.544700089992967,
+      "grad_norm": 26.045589447021484,
+      "learning_rate": 1.5455317214605142e-05,
+      "loss": 0.4403,
+      "step": 369920
+    },
+    {
+      "epoch": 4.547845236669216,
+      "grad_norm": 0.873694658279419,
+      "learning_rate": 1.545217209690894e-05,
+      "loss": 0.4208,
+      "step": 370176
+    },
+    {
+      "epoch": 4.550990383345465,
+      "grad_norm": 45.83314895629883,
+      "learning_rate": 1.544902697921274e-05,
+      "loss": 0.4697,
+      "step": 370432
+    },
+    {
+      "epoch": 4.554135530021715,
+      "grad_norm": 10.581768035888672,
+      "learning_rate": 1.544588186151654e-05,
+      "loss": 0.4825,
+      "step": 370688
+    },
+    {
+      "epoch": 4.557280676697965,
+      "grad_norm": 0.19337469339370728,
+      "learning_rate": 1.5442736743820338e-05,
+      "loss": 0.4637,
+      "step": 370944
+    },
+    {
+      "epoch": 4.560425823374214,
+      "grad_norm": 13.458893775939941,
+      "learning_rate": 1.5439591626124135e-05,
+      "loss": 0.3764,
+      "step": 371200
+    },
+    {
+      "epoch": 4.563570970050463,
+      "grad_norm": 45.62543487548828,
+      "learning_rate": 1.5436446508427934e-05,
+      "loss": 0.4887,
+      "step": 371456
+    },
+    {
+      "epoch": 4.566716116726713,
+      "grad_norm": 0.36106231808662415,
+      "learning_rate": 1.543330139073173e-05,
+      "loss": 0.3956,
+      "step": 371712
+    },
+    {
+      "epoch": 4.569861263402963,
+      "grad_norm": 43.74862289428711,
+      "learning_rate": 1.543015627303553e-05,
+      "loss": 0.5349,
+      "step": 371968
+    },
+    {
+      "epoch": 4.573006410079213,
+      "grad_norm": 29.026458740234375,
+      "learning_rate": 1.542701115533933e-05,
+      "loss": 0.425,
+      "step": 372224
+    },
+    {
+      "epoch": 4.576151556755462,
+      "grad_norm": 81.37934112548828,
+      "learning_rate": 1.542386603764313e-05,
+      "loss": 0.5009,
+      "step": 372480
+    },
+    {
+      "epoch": 4.579296703431711,
+      "grad_norm": 0.5218445658683777,
+      "learning_rate": 1.5420720919946927e-05,
+      "loss": 0.4872,
+      "step": 372736
+    },
+    {
+      "epoch": 4.582441850107961,
+      "grad_norm": 1.298308253288269,
+      "learning_rate": 1.5417575802250727e-05,
+      "loss": 0.4024,
+      "step": 372992
+    },
+    {
+      "epoch": 4.58558699678421,
+      "grad_norm": 1.760785698890686,
+      "learning_rate": 1.5414430684554523e-05,
+      "loss": 0.4659,
+      "step": 373248
+    },
+    {
+      "epoch": 4.58873214346046,
+      "grad_norm": 0.6107638478279114,
+      "learning_rate": 1.5411285566858323e-05,
+      "loss": 0.4418,
+      "step": 373504
+    },
+    {
+      "epoch": 4.5918772901367095,
+      "grad_norm": 313.6508483886719,
+      "learning_rate": 1.5408140449162123e-05,
+      "loss": 0.4364,
+      "step": 373760
+    },
+    {
+      "epoch": 4.595022436812959,
+      "grad_norm": 29.36980628967285,
+      "learning_rate": 1.5404995331465922e-05,
+      "loss": 0.4375,
+      "step": 374016
+    },
+    {
+      "epoch": 4.598167583489208,
+      "grad_norm": 33.21347427368164,
+      "learning_rate": 1.540185021376972e-05,
+      "loss": 0.4225,
+      "step": 374272
+    },
+    {
+      "epoch": 4.601312730165458,
+      "grad_norm": 1.268494963645935,
+      "learning_rate": 1.539870509607352e-05,
+      "loss": 0.5491,
+      "step": 374528
+    },
+    {
+      "epoch": 4.604457876841708,
+      "grad_norm": 0.4023263156414032,
+      "learning_rate": 1.539555997837732e-05,
+      "loss": 0.4278,
+      "step": 374784
+    },
+    {
+      "epoch": 4.6076030235179575,
+      "grad_norm": 90.2680435180664,
+      "learning_rate": 1.5392414860681115e-05,
+      "loss": 0.4145,
+      "step": 375040
+    },
+    {
+      "epoch": 4.610748170194206,
+      "grad_norm": 0.36424434185028076,
+      "learning_rate": 1.5389269742984915e-05,
+      "loss": 0.3873,
+      "step": 375296
+    },
+    {
+      "epoch": 4.613893316870456,
+      "grad_norm": 76.73234558105469,
+      "learning_rate": 1.5386124625288714e-05,
+      "loss": 0.502,
+      "step": 375552
+    },
+    {
+      "epoch": 4.617038463546706,
+      "grad_norm": 0.16531279683113098,
+      "learning_rate": 1.5382979507592514e-05,
+      "loss": 0.4933,
+      "step": 375808
+    },
+    {
+      "epoch": 4.620183610222956,
+      "grad_norm": 54.908939361572266,
+      "learning_rate": 1.537983438989631e-05,
+      "loss": 0.3921,
+      "step": 376064
+    },
+    {
+      "epoch": 4.6233287568992045,
+      "grad_norm": 32.6933708190918,
+      "learning_rate": 1.537668927220011e-05,
+      "loss": 0.4481,
+      "step": 376320
+    },
+    {
+      "epoch": 4.626473903575454,
+      "grad_norm": 0.17195437848567963,
+      "learning_rate": 1.5373544154503907e-05,
+      "loss": 0.3818,
+      "step": 376576
+    },
+    {
+      "epoch": 4.629619050251704,
+      "grad_norm": 0.4807603061199188,
+      "learning_rate": 1.5370399036807707e-05,
+      "loss": 0.4113,
+      "step": 376832
+    },
+    {
+      "epoch": 4.632764196927953,
+      "grad_norm": 0.3930334448814392,
+      "learning_rate": 1.5367253919111503e-05,
+      "loss": 0.4807,
+      "step": 377088
+    },
+    {
+      "epoch": 4.635909343604203,
+      "grad_norm": 30.09731674194336,
+      "learning_rate": 1.5364108801415306e-05,
+      "loss": 0.5704,
+      "step": 377344
+    },
+    {
+      "epoch": 4.6390544902804525,
+      "grad_norm": 40.934322357177734,
+      "learning_rate": 1.5360963683719103e-05,
+      "loss": 0.4741,
+      "step": 377600
+    },
+    {
+      "epoch": 4.642199636956702,
+      "grad_norm": 36.52864456176758,
+      "learning_rate": 1.5357818566022902e-05,
+      "loss": 0.2625,
+      "step": 377856
+    },
+    {
+      "epoch": 4.645344783632952,
+      "grad_norm": 123.2453842163086,
+      "learning_rate": 1.53546734483267e-05,
+      "loss": 0.4922,
+      "step": 378112
+    },
+    {
+      "epoch": 4.648489930309201,
+      "grad_norm": 0.5196950435638428,
+      "learning_rate": 1.53515283306305e-05,
+      "loss": 0.5214,
+      "step": 378368
+    },
+    {
+      "epoch": 4.651635076985451,
+      "grad_norm": 12.661064147949219,
+      "learning_rate": 1.53483832129343e-05,
+      "loss": 0.4796,
+      "step": 378624
+    },
+    {
+      "epoch": 4.6547802236617,
+      "grad_norm": 11.640763282775879,
+      "learning_rate": 1.5345238095238098e-05,
+      "loss": 0.4067,
+      "step": 378880
+    },
+    {
+      "epoch": 4.657925370337949,
+      "grad_norm": 8.130566596984863,
+      "learning_rate": 1.5342092977541895e-05,
+      "loss": 0.5246,
+      "step": 379136
+    },
+    {
+      "epoch": 4.661070517014199,
+      "grad_norm": 0.15848112106323242,
+      "learning_rate": 1.5338947859845695e-05,
+      "loss": 0.3841,
+      "step": 379392
+    },
+    {
+      "epoch": 4.664215663690449,
+      "grad_norm": 0.3735288083553314,
+      "learning_rate": 1.533580274214949e-05,
+      "loss": 0.538,
+      "step": 379648
+    },
+    {
+      "epoch": 4.667360810366699,
+      "grad_norm": 0.5919455885887146,
+      "learning_rate": 1.533265762445329e-05,
+      "loss": 0.3892,
+      "step": 379904
+    },
+    {
+      "epoch": 4.6705059570429475,
+      "grad_norm": 205.1002655029297,
+      "learning_rate": 1.532951250675709e-05,
+      "loss": 0.3532,
+      "step": 380160
+    },
+    {
+      "epoch": 4.673651103719197,
+      "grad_norm": 5.330204010009766,
+      "learning_rate": 1.5326367389060887e-05,
+      "loss": 0.4009,
+      "step": 380416
+    },
+    {
+      "epoch": 4.676796250395447,
+      "grad_norm": 2.469081401824951,
+      "learning_rate": 1.5323222271364687e-05,
+      "loss": 0.5124,
+      "step": 380672
+    },
+    {
+      "epoch": 4.679941397071697,
+      "grad_norm": 0.9177038669586182,
+      "learning_rate": 1.5320077153668487e-05,
+      "loss": 0.5145,
+      "step": 380928
+    },
+    {
+      "epoch": 4.683086543747946,
+      "grad_norm": 0.021899504587054253,
+      "learning_rate": 1.5316932035972286e-05,
+      "loss": 0.4938,
+      "step": 381184
+    },
+    {
+      "epoch": 4.686231690424195,
+      "grad_norm": 5.4605255126953125,
+      "learning_rate": 1.5313786918276083e-05,
+      "loss": 0.4124,
+      "step": 381440
+    },
+    {
+      "epoch": 4.689376837100445,
+      "grad_norm": 36.440486907958984,
+      "learning_rate": 1.5310641800579883e-05,
+      "loss": 0.4617,
+      "step": 381696
+    },
+    {
+      "epoch": 4.692521983776695,
+      "grad_norm": 0.18164685368537903,
+      "learning_rate": 1.530749668288368e-05,
+      "loss": 0.436,
+      "step": 381952
+    },
+    {
+      "epoch": 4.695667130452944,
+      "grad_norm": 1.0528472661972046,
+      "learning_rate": 1.5304351565187482e-05,
+      "loss": 0.4793,
+      "step": 382208
+    },
+    {
+      "epoch": 4.698812277129194,
+      "grad_norm": 18.716480255126953,
+      "learning_rate": 1.530120644749128e-05,
+      "loss": 0.4567,
+      "step": 382464
+    },
+    {
+      "epoch": 4.701957423805443,
+      "grad_norm": 0.9601308107376099,
+      "learning_rate": 1.529806132979508e-05,
+      "loss": 0.4811,
+      "step": 382720
+    },
+    {
+      "epoch": 4.705102570481693,
+      "grad_norm": 0.24384582042694092,
+      "learning_rate": 1.5294916212098875e-05,
+      "loss": 0.3727,
+      "step": 382976
+    },
+    {
+      "epoch": 4.708247717157942,
+      "grad_norm": 63.768585205078125,
+      "learning_rate": 1.5291771094402675e-05,
+      "loss": 0.457,
+      "step": 383232
+    },
+    {
+      "epoch": 4.711392863834192,
+      "grad_norm": 33.049888610839844,
+      "learning_rate": 1.528862597670647e-05,
+      "loss": 0.4236,
+      "step": 383488
+    },
+    {
+      "epoch": 4.7145380105104415,
+      "grad_norm": 37.922874450683594,
+      "learning_rate": 1.528548085901027e-05,
+      "loss": 0.4611,
+      "step": 383744
+    },
+    {
+      "epoch": 4.717683157186691,
+      "grad_norm": 15.72169303894043,
+      "learning_rate": 1.528233574131407e-05,
+      "loss": 0.4542,
+      "step": 384000
+    },
+    {
+      "epoch": 4.72082830386294,
+      "grad_norm": 42.17428970336914,
+      "learning_rate": 1.527919062361787e-05,
+      "loss": 0.4517,
+      "step": 384256
+    },
+    {
+      "epoch": 4.72397345053919,
+      "grad_norm": 420.9949645996094,
+      "learning_rate": 1.5276045505921667e-05,
+      "loss": 0.4932,
+      "step": 384512
+    },
+    {
+      "epoch": 4.72711859721544,
+      "grad_norm": 0.014308485202491283,
+      "learning_rate": 1.5272900388225467e-05,
+      "loss": 0.4584,
+      "step": 384768
+    },
+    {
+      "epoch": 4.730263743891689,
+      "grad_norm": 122.4588394165039,
+      "learning_rate": 1.5269755270529266e-05,
+      "loss": 0.3775,
+      "step": 385024
+    },
+    {
+      "epoch": 4.733408890567938,
+      "grad_norm": 22.24554443359375,
+      "learning_rate": 1.5266610152833063e-05,
+      "loss": 0.4685,
+      "step": 385280
+    },
+    {
+      "epoch": 4.736554037244188,
+      "grad_norm": 5.365596771240234,
+      "learning_rate": 1.5263465035136863e-05,
+      "loss": 0.4637,
+      "step": 385536
+    },
+    {
+      "epoch": 4.739699183920438,
+      "grad_norm": 75.9616470336914,
+      "learning_rate": 1.5260319917440663e-05,
+      "loss": 0.4782,
+      "step": 385792
+    },
+    {
+      "epoch": 4.742844330596688,
+      "grad_norm": 78.20529174804688,
+      "learning_rate": 1.525717479974446e-05,
+      "loss": 0.4626,
+      "step": 386048
+    },
+    {
+      "epoch": 4.7459894772729365,
+      "grad_norm": 0.11991903185844421,
+      "learning_rate": 1.5254029682048259e-05,
+      "loss": 0.4615,
+      "step": 386304
+    },
+    {
+      "epoch": 4.749134623949186,
+      "grad_norm": 0.5266178846359253,
+      "learning_rate": 1.5250884564352057e-05,
+      "loss": 0.4312,
+      "step": 386560
+    },
+    {
+      "epoch": 4.752279770625436,
+      "grad_norm": 0.10850125551223755,
+      "learning_rate": 1.5247739446655857e-05,
+      "loss": 0.5046,
+      "step": 386816
+    },
+    {
+      "epoch": 4.755424917301685,
+      "grad_norm": 0.3127959072589874,
+      "learning_rate": 1.5244594328959655e-05,
+      "loss": 0.3982,
+      "step": 387072
+    },
+    {
+      "epoch": 4.758570063977935,
+      "grad_norm": 12.409524917602539,
+      "learning_rate": 1.5241449211263455e-05,
+      "loss": 0.5496,
+      "step": 387328
+    },
+    {
+      "epoch": 4.7617152106541845,
+      "grad_norm": 4.042536735534668,
+      "learning_rate": 1.5238304093567253e-05,
+      "loss": 0.4297,
+      "step": 387584
+    },
+    {
+      "epoch": 4.764860357330434,
+      "grad_norm": 24.607297897338867,
+      "learning_rate": 1.5235158975871052e-05,
+      "loss": 0.4218,
+      "step": 387840
+    },
+    {
+      "epoch": 4.768005504006683,
+      "grad_norm": 74.20430755615234,
+      "learning_rate": 1.523201385817485e-05,
+      "loss": 0.5331,
+      "step": 388096
+    },
+    {
+      "epoch": 4.771150650682933,
+      "grad_norm": 93.53345489501953,
+      "learning_rate": 1.5228868740478649e-05,
+      "loss": 0.4141,
+      "step": 388352
+    },
+    {
+      "epoch": 4.774295797359183,
+      "grad_norm": 10.38814926147461,
+      "learning_rate": 1.5225723622782447e-05,
+      "loss": 0.45,
+      "step": 388608
+    },
+    {
+      "epoch": 4.777440944035432,
+      "grad_norm": 71.80351257324219,
+      "learning_rate": 1.5222578505086247e-05,
+      "loss": 0.4434,
+      "step": 388864
+    },
+    {
+      "epoch": 4.780586090711681,
+      "grad_norm": 0.0529104620218277,
+      "learning_rate": 1.5219433387390046e-05,
+      "loss": 0.3809,
+      "step": 389120
+    },
+    {
+      "epoch": 4.783731237387931,
+      "grad_norm": 47.54201889038086,
+      "learning_rate": 1.5216288269693845e-05,
+      "loss": 0.4128,
+      "step": 389376
+    },
+    {
+      "epoch": 4.786876384064181,
+      "grad_norm": 8.440589904785156,
+      "learning_rate": 1.5213143151997643e-05,
+      "loss": 0.4178,
+      "step": 389632
+    },
+    {
+      "epoch": 4.790021530740431,
+      "grad_norm": 0.1548546403646469,
+      "learning_rate": 1.520999803430144e-05,
+      "loss": 0.4209,
+      "step": 389888
+    },
+    {
+      "epoch": 4.7931666774166795,
+      "grad_norm": 89.0304183959961,
+      "learning_rate": 1.5206852916605239e-05,
+      "loss": 0.4189,
+      "step": 390144
+    },
+    {
+      "epoch": 4.796311824092929,
+      "grad_norm": 42.92784118652344,
+      "learning_rate": 1.5203707798909037e-05,
+      "loss": 0.4871,
+      "step": 390400
+    },
+    {
+      "epoch": 4.799456970769179,
+      "grad_norm": 248.5048828125,
+      "learning_rate": 1.5200562681212838e-05,
+      "loss": 0.4569,
+      "step": 390656
+    },
+    {
+      "epoch": 4.802602117445428,
+      "grad_norm": 7.337618827819824,
+      "learning_rate": 1.5197417563516637e-05,
+      "loss": 0.4461,
+      "step": 390912
+    },
+    {
+      "epoch": 4.805747264121678,
+      "grad_norm": 0.14882519841194153,
+      "learning_rate": 1.5194272445820435e-05,
+      "loss": 0.422,
+      "step": 391168
+    },
+    {
+      "epoch": 4.808892410797927,
+      "grad_norm": 0.7318896055221558,
+      "learning_rate": 1.5191127328124233e-05,
+      "loss": 0.4024,
+      "step": 391424
+    },
+    {
+      "epoch": 4.812037557474177,
+      "grad_norm": 78.16159057617188,
+      "learning_rate": 1.5187982210428031e-05,
+      "loss": 0.5289,
+      "step": 391680
+    },
+    {
+      "epoch": 4.815182704150427,
+      "grad_norm": 21.80522346496582,
+      "learning_rate": 1.518483709273183e-05,
+      "loss": 0.4007,
+      "step": 391936
+    },
+    {
+      "epoch": 4.818327850826676,
+      "grad_norm": 76.2160873413086,
+      "learning_rate": 1.518169197503563e-05,
+      "loss": 0.4581,
+      "step": 392192
+    },
+    {
+      "epoch": 4.821472997502926,
+      "grad_norm": 2.214731216430664,
+      "learning_rate": 1.5178546857339429e-05,
+      "loss": 0.4541,
+      "step": 392448
+    },
+    {
+      "epoch": 4.824618144179175,
+      "grad_norm": 5.559924125671387,
+      "learning_rate": 1.5175401739643227e-05,
+      "loss": 0.4577,
+      "step": 392704
+    },
+    {
+      "epoch": 4.827763290855424,
+      "grad_norm": 32.65327835083008,
+      "learning_rate": 1.5172256621947027e-05,
+      "loss": 0.4104,
+      "step": 392960
+    },
+    {
+      "epoch": 4.830908437531674,
+      "grad_norm": 4.551636695861816,
+      "learning_rate": 1.5169111504250825e-05,
+      "loss": 0.3861,
+      "step": 393216
+    },
+    {
+      "epoch": 4.834053584207924,
+      "grad_norm": 27.39976692199707,
+      "learning_rate": 1.5165966386554623e-05,
+      "loss": 0.4473,
+      "step": 393472
+    },
+    {
+      "epoch": 4.8371987308841735,
+      "grad_norm": 0.37570255994796753,
+      "learning_rate": 1.5162821268858421e-05,
+      "loss": 0.437,
+      "step": 393728
+    },
+    {
+      "epoch": 4.840343877560422,
+      "grad_norm": 0.3768576979637146,
+      "learning_rate": 1.515967615116222e-05,
+      "loss": 0.4507,
+      "step": 393984
+    },
+    {
+      "epoch": 4.843489024236672,
+      "grad_norm": 64.30237579345703,
+      "learning_rate": 1.515653103346602e-05,
+      "loss": 0.3845,
+      "step": 394240
+    },
+    {
+      "epoch": 4.846634170912922,
+      "grad_norm": 1.057909607887268,
+      "learning_rate": 1.5153385915769819e-05,
+      "loss": 0.5416,
+      "step": 394496
+    },
+    {
+      "epoch": 4.849779317589172,
+      "grad_norm": 0.27367573976516724,
+      "learning_rate": 1.5150240798073617e-05,
+      "loss": 0.4589,
+      "step": 394752
+    },
+    {
+      "epoch": 4.852924464265421,
+      "grad_norm": 32.440677642822266,
+      "learning_rate": 1.5147095680377415e-05,
+      "loss": 0.4443,
+      "step": 395008
+    },
+    {
+      "epoch": 4.85606961094167,
+      "grad_norm": 21.873031616210938,
+      "learning_rate": 1.5143950562681213e-05,
+      "loss": 0.5031,
+      "step": 395264
+    },
+    {
+      "epoch": 4.85921475761792,
+      "grad_norm": 0.22820694744586945,
+      "learning_rate": 1.5140805444985011e-05,
+      "loss": 0.479,
+      "step": 395520
+    },
+    {
+      "epoch": 4.86235990429417,
+      "grad_norm": 40.81998062133789,
+      "learning_rate": 1.5137660327288813e-05,
+      "loss": 0.5438,
+      "step": 395776
+    },
+    {
+      "epoch": 4.865505050970419,
+      "grad_norm": 96.40206146240234,
+      "learning_rate": 1.513451520959261e-05,
+      "loss": 0.3843,
+      "step": 396032
+    },
+    {
+      "epoch": 4.8686501976466685,
+      "grad_norm": 92.08551025390625,
+      "learning_rate": 1.5131370091896409e-05,
+      "loss": 0.3889,
+      "step": 396288
+    },
+    {
+      "epoch": 4.871795344322918,
+      "grad_norm": 0.3882986307144165,
+      "learning_rate": 1.5128224974200207e-05,
+      "loss": 0.4773,
+      "step": 396544
+    },
+    {
+      "epoch": 4.874940490999167,
+      "grad_norm": 0.09950807690620422,
+      "learning_rate": 1.5125079856504005e-05,
+      "loss": 0.4041,
+      "step": 396800
+    },
+    {
+      "epoch": 4.878085637675417,
+      "grad_norm": 43.67584991455078,
+      "learning_rate": 1.5121934738807805e-05,
+      "loss": 0.443,
+      "step": 397056
+    },
+    {
+      "epoch": 4.881230784351667,
+      "grad_norm": 3.2849390506744385,
+      "learning_rate": 1.5118789621111605e-05,
+      "loss": 0.4526,
+      "step": 397312
+    },
+    {
+      "epoch": 4.8843759310279165,
+      "grad_norm": 8.674638748168945,
+      "learning_rate": 1.5115644503415403e-05,
+      "loss": 0.4378,
+      "step": 397568
+    },
+    {
+      "epoch": 4.887521077704166,
+      "grad_norm": 40.49319839477539,
+      "learning_rate": 1.51124993857192e-05,
+      "loss": 0.4479,
+      "step": 397824
+    },
+    {
+      "epoch": 4.890666224380415,
+      "grad_norm": 4.415699005126953,
+      "learning_rate": 1.5109354268023e-05,
+      "loss": 0.4844,
+      "step": 398080
+    },
+    {
+      "epoch": 4.893811371056665,
+      "grad_norm": 0.09496615082025528,
+      "learning_rate": 1.5106209150326799e-05,
+      "loss": 0.424,
+      "step": 398336
+    },
+    {
+      "epoch": 4.896956517732915,
+      "grad_norm": 1.7374989986419678,
+      "learning_rate": 1.5103064032630597e-05,
+      "loss": 0.4511,
+      "step": 398592
+    },
+    {
+      "epoch": 4.9001016644091635,
+      "grad_norm": 21.716785430908203,
+      "learning_rate": 1.5099918914934395e-05,
+      "loss": 0.3953,
+      "step": 398848
+    },
+    {
+      "epoch": 4.903246811085413,
+      "grad_norm": 100.94387817382812,
+      "learning_rate": 1.5096773797238196e-05,
+      "loss": 0.4831,
+      "step": 399104
+    },
+    {
+      "epoch": 4.906391957761663,
+      "grad_norm": 0.1464122086763382,
+      "learning_rate": 1.5093628679541995e-05,
+      "loss": 0.443,
+      "step": 399360
+    },
+    {
+      "epoch": 4.909537104437913,
+      "grad_norm": 19.914461135864258,
+      "learning_rate": 1.5090483561845793e-05,
+      "loss": 0.4881,
+      "step": 399616
+    },
+    {
+      "epoch": 4.912682251114162,
+      "grad_norm": 81.96047973632812,
+      "learning_rate": 1.508733844414959e-05,
+      "loss": 0.4267,
+      "step": 399872
+    },
+    {
+      "epoch": 4.9158273977904114,
+      "grad_norm": 123.0934829711914,
+      "learning_rate": 1.5084193326453389e-05,
+      "loss": 0.4732,
+      "step": 400128
+    },
+    {
+      "epoch": 4.918972544466661,
+      "grad_norm": 0.3333633244037628,
+      "learning_rate": 1.5081048208757187e-05,
+      "loss": 0.4834,
+      "step": 400384
+    },
+    {
+      "epoch": 4.922117691142911,
+      "grad_norm": 1.244377613067627,
+      "learning_rate": 1.5077903091060988e-05,
+      "loss": 0.4714,
+      "step": 400640
+    },
+    {
+      "epoch": 4.92526283781916,
+      "grad_norm": 0.4310603737831116,
+      "learning_rate": 1.5074757973364787e-05,
+      "loss": 0.4822,
+      "step": 400896
+    },
+    {
+      "epoch": 4.92840798449541,
+      "grad_norm": 0.4027729034423828,
+      "learning_rate": 1.5071612855668585e-05,
+      "loss": 0.4321,
+      "step": 401152
+    },
+    {
+      "epoch": 4.931553131171659,
+      "grad_norm": 13.537920951843262,
+      "learning_rate": 1.5068467737972383e-05,
+      "loss": 0.5037,
+      "step": 401408
+    },
+    {
+      "epoch": 4.934698277847909,
+      "grad_norm": 83.20620727539062,
+      "learning_rate": 1.5065322620276181e-05,
+      "loss": 0.4267,
+      "step": 401664
+    },
+    {
+      "epoch": 4.937843424524158,
+      "grad_norm": 104.27676391601562,
+      "learning_rate": 1.506217750257998e-05,
+      "loss": 0.4385,
+      "step": 401920
+    },
+    {
+      "epoch": 4.940988571200408,
+      "grad_norm": 0.3629232347011566,
+      "learning_rate": 1.5059032384883779e-05,
+      "loss": 0.4539,
+      "step": 402176
+    },
+    {
+      "epoch": 4.944133717876658,
+      "grad_norm": 26.156049728393555,
+      "learning_rate": 1.5055887267187579e-05,
+      "loss": 0.5129,
+      "step": 402432
+    },
+    {
+      "epoch": 4.947278864552907,
+      "grad_norm": 11.355233192443848,
+      "learning_rate": 1.5052742149491377e-05,
+      "loss": 0.5034,
+      "step": 402688
+    },
+    {
+      "epoch": 4.950424011229156,
+      "grad_norm": 0.15290385484695435,
+      "learning_rate": 1.5049597031795175e-05,
+      "loss": 0.4437,
+      "step": 402944
+    },
+    {
+      "epoch": 4.953569157905406,
+      "grad_norm": 0.03169036656618118,
+      "learning_rate": 1.5046451914098975e-05,
+      "loss": 0.5498,
+      "step": 403200
+    },
+    {
+      "epoch": 4.956714304581656,
+      "grad_norm": 0.6258155107498169,
+      "learning_rate": 1.5043306796402773e-05,
+      "loss": 0.3878,
+      "step": 403456
+    },
+    {
+      "epoch": 4.9598594512579055,
+      "grad_norm": 42.20677947998047,
+      "learning_rate": 1.5040161678706571e-05,
+      "loss": 0.5044,
+      "step": 403712
+    },
+    {
+      "epoch": 4.963004597934154,
+      "grad_norm": 0.22367466986179352,
+      "learning_rate": 1.503701656101037e-05,
+      "loss": 0.4911,
+      "step": 403968
+    },
+    {
+      "epoch": 4.966149744610404,
+      "grad_norm": 1.793988823890686,
+      "learning_rate": 1.503387144331417e-05,
+      "loss": 0.3491,
+      "step": 404224
+    },
+    {
+      "epoch": 4.969294891286654,
+      "grad_norm": 2.8468687534332275,
+      "learning_rate": 1.5030726325617969e-05,
+      "loss": 0.4808,
+      "step": 404480
+    },
+    {
+      "epoch": 4.972440037962903,
+      "grad_norm": 55.412418365478516,
+      "learning_rate": 1.5027581207921767e-05,
+      "loss": 0.4899,
+      "step": 404736
+    },
+    {
+      "epoch": 4.975585184639153,
+      "grad_norm": 0.14198999106884003,
+      "learning_rate": 1.5024436090225565e-05,
+      "loss": 0.4653,
+      "step": 404992
+    },
+    {
+      "epoch": 4.978730331315402,
+      "grad_norm": 2.9037771224975586,
+      "learning_rate": 1.5021290972529363e-05,
+      "loss": 0.4385,
+      "step": 405248
+    },
+    {
+      "epoch": 4.981875477991652,
+      "grad_norm": 14.98426342010498,
+      "learning_rate": 1.5018145854833161e-05,
+      "loss": 0.4587,
+      "step": 405504
+    },
+    {
+      "epoch": 4.985020624667902,
+      "grad_norm": 15.498087882995605,
+      "learning_rate": 1.5015000737136963e-05,
+      "loss": 0.4419,
+      "step": 405760
+    },
+    {
+      "epoch": 4.988165771344151,
+      "grad_norm": 5.2790374755859375,
+      "learning_rate": 1.501185561944076e-05,
+      "loss": 0.5088,
+      "step": 406016
+    },
+    {
+      "epoch": 4.9913109180204005,
+      "grad_norm": 197.16395568847656,
+      "learning_rate": 1.5008710501744559e-05,
+      "loss": 0.532,
+      "step": 406272
+    },
+    {
+      "epoch": 4.99445606469665,
+      "grad_norm": 10.321364402770996,
+      "learning_rate": 1.5005565384048357e-05,
+      "loss": 0.5187,
+      "step": 406528
+    },
+    {
+      "epoch": 4.997601211372899,
+      "grad_norm": 0.1803307980298996,
+      "learning_rate": 1.5002420266352155e-05,
+      "loss": 0.4847,
+      "step": 406784
+    },
+    {
+      "epoch": 5.0,
+      "eval_AGE": {
+        "f1": 0.9686648918265554,
+        "number": 32192,
+        "precision": 0.9544412953024182,
+        "recall": 0.9833188369781312
+      },
+      "eval_BUILDINGNUM": {
+        "f1": 0.980485793795307,
+        "number": 28889,
+        "precision": 0.9731969309462916,
+        "recall": 0.9878846619820693
+      },
+      "eval_CITY": {
+        "f1": 0.9723731721068605,
+        "number": 41669,
+        "precision": 0.972104960184208,
+        "recall": 0.9726415320742039
+      },
+      "eval_DATE": {
+        "f1": 0.997593216573481,
+        "number": 80940,
+        "precision": 0.9965968780054746,
+        "recall": 0.9985915492957746
+      },
+      "eval_EMAIL": {
+        "f1": 0.9970628853747253,
+        "number": 51445,
+        "precision": 0.9977227163906028,
+        "recall": 0.9964039265234717
+      },
+      "eval_GENDER": {
+        "f1": 0.9728994151369506,
+        "number": 17457,
+        "precision": 0.9836075171242902,
+        "recall": 0.9624219510797961
+      },
+      "eval_GIVENNAME": {
+        "f1": 0.9333472131342573,
+        "number": 100350,
+        "precision": 0.9285967923299994,
+        "recall": 0.9381464872944694
+      },
+      "eval_IDCARDNUM": {
+        "f1": 0.9075509805199037,
+        "number": 22068,
+        "precision": 0.8713874640310272,
+        "recall": 0.9468461120174008
+      },
+      "eval_PASSPORTNUM": {
+        "f1": 0.9233025823136605,
+        "number": 12959,
+        "precision": 0.9236947791164659,
+        "recall": 0.9229107184196311
+      },
+      "eval_SEX": {
+        "f1": 0.9661364629922318,
+        "number": 14570,
+        "precision": 0.9471220412738182,
+        "recall": 0.985929993136582
+      },
+      "eval_SOCIALNUM": {
+        "f1": 0.9101440392277046,
+        "number": 15327,
+        "precision": 0.8581748829682714,
+        "recall": 0.9688132054544268
+      },
+      "eval_STREET": {
+        "f1": 0.9766071266660622,
+        "number": 32803,
+        "precision": 0.9682958257169398,
+        "recall": 0.9850623418589763
+      },
+      "eval_SURNAME": {
+        "f1": 0.9277010560519904,
+        "number": 85192,
+        "precision": 0.9304725692558392,
+        "recall": 0.9249460043196545
+      },
+      "eval_TELEPHONENUM": {
+        "f1": 0.9934911162532646,
+        "number": 36941,
+        "precision": 0.993262622436279,
+        "recall": 0.9937197152215695
+      },
+      "eval_ZIPCODE": {
+        "f1": 0.9835614186495706,
+        "number": 24346,
+        "precision": 0.9840871710526315,
+        "recall": 0.9830362277170788
+      },
+      "eval_overall_accuracy": 0.9931339995806231,
+      "eval_overall_f1": 0.961819476188692,
+      "eval_overall_precision": 0.9564764505503787,
+      "eval_overall_recall": 0.9672225310978183,
+      "eval_runtime": 3726.8328,
+      "eval_samples_per_second": 87.362,
+      "eval_steps_per_second": 43.681,
+      "step": 406980
+    },
+    {
+      "epoch": 5.000737143752246,
+      "grad_norm": 0.7710248827934265,
+      "learning_rate": 1.4999275148655955e-05,
+      "loss": 0.4059,
+      "step": 407040
+    },
+    {
+      "epoch": 5.003882290428495,
+      "grad_norm": 0.09372542053461075,
+      "learning_rate": 1.4996130030959755e-05,
+      "loss": 0.4181,
+      "step": 407296
+    },
+    {
+      "epoch": 5.007027437104745,
+      "grad_norm": 1.1031941175460815,
+      "learning_rate": 1.4992984913263553e-05,
+      "loss": 0.2997,
+      "step": 407552
+    },
+    {
+      "epoch": 5.010172583780995,
+      "grad_norm": 12.962038040161133,
+      "learning_rate": 1.498983979556735e-05,
+      "loss": 0.401,
+      "step": 407808
+    },
+    {
+      "epoch": 5.013317730457244,
+      "grad_norm": 0.2239060252904892,
+      "learning_rate": 1.4986694677871149e-05,
+      "loss": 0.3475,
+      "step": 408064
+    },
+    {
+      "epoch": 5.016462877133494,
+      "grad_norm": 0.25520265102386475,
+      "learning_rate": 1.4983549560174949e-05,
+      "loss": 0.3334,
+      "step": 408320
+    },
+    {
+      "epoch": 5.019608023809743,
+      "grad_norm": 2.983374834060669,
+      "learning_rate": 1.4980404442478747e-05,
+      "loss": 0.3308,
+      "step": 408576
+    },
+    {
+      "epoch": 5.022753170485993,
+      "grad_norm": 0.2632104456424713,
+      "learning_rate": 1.4977259324782545e-05,
+      "loss": 0.3954,
+      "step": 408832
+    },
+    {
+      "epoch": 5.025898317162242,
+      "grad_norm": 35.5607795715332,
+      "learning_rate": 1.4974114207086345e-05,
+      "loss": 0.3856,
+      "step": 409088
+    },
+    {
+      "epoch": 5.029043463838492,
+      "grad_norm": 0.08944360166788101,
+      "learning_rate": 1.4970969089390145e-05,
+      "loss": 0.2983,
+      "step": 409344
+    },
+    {
+      "epoch": 5.0321886105147415,
+      "grad_norm": 34.97138214111328,
+      "learning_rate": 1.4967823971693943e-05,
+      "loss": 0.3283,
+      "step": 409600
+    },
+    {
+      "epoch": 5.035333757190991,
+      "grad_norm": 5.585193634033203,
+      "learning_rate": 1.496467885399774e-05,
+      "loss": 0.3455,
+      "step": 409856
+    },
+    {
+      "epoch": 5.03847890386724,
+      "grad_norm": 1.1885769367218018,
+      "learning_rate": 1.4961533736301539e-05,
+      "loss": 0.4045,
+      "step": 410112
+    },
+    {
+      "epoch": 5.04162405054349,
+      "grad_norm": 20.224523544311523,
+      "learning_rate": 1.4958388618605337e-05,
+      "loss": 0.3256,
+      "step": 410368
+    },
+    {
+      "epoch": 5.04476919721974,
+      "grad_norm": 41.984886169433594,
+      "learning_rate": 1.4955243500909138e-05,
+      "loss": 0.3389,
+      "step": 410624
+    },
+    {
+      "epoch": 5.047914343895989,
+      "grad_norm": 0.1993991881608963,
+      "learning_rate": 1.4952098383212937e-05,
+      "loss": 0.355,
+      "step": 410880
+    },
+    {
+      "epoch": 5.051059490572238,
+      "grad_norm": 0.35691314935684204,
+      "learning_rate": 1.4948953265516735e-05,
+      "loss": 0.2699,
+      "step": 411136
+    },
+    {
+      "epoch": 5.054204637248488,
+      "grad_norm": 0.44386595487594604,
+      "learning_rate": 1.4945808147820533e-05,
+      "loss": 0.2916,
+      "step": 411392
+    },
+    {
+      "epoch": 5.057349783924738,
+      "grad_norm": 107.46903228759766,
+      "learning_rate": 1.4942663030124331e-05,
+      "loss": 0.3219,
+      "step": 411648
+    },
+    {
+      "epoch": 5.060494930600987,
+      "grad_norm": 0.08264636993408203,
+      "learning_rate": 1.4939517912428129e-05,
+      "loss": 0.3817,
+      "step": 411904
+    },
+    {
+      "epoch": 5.0636400772772365,
+      "grad_norm": 6.605735778808594,
+      "learning_rate": 1.4936372794731929e-05,
+      "loss": 0.3802,
+      "step": 412160
+    },
+    {
+      "epoch": 5.066785223953486,
+      "grad_norm": 0.5865673422813416,
+      "learning_rate": 1.4933227677035729e-05,
+      "loss": 0.4049,
+      "step": 412416
+    },
+    {
+      "epoch": 5.069930370629736,
+      "grad_norm": 0.06085227429866791,
+      "learning_rate": 1.4930082559339527e-05,
+      "loss": 0.3209,
+      "step": 412672
+    },
+    {
+      "epoch": 5.073075517305985,
+      "grad_norm": 0.6298277378082275,
+      "learning_rate": 1.4926937441643325e-05,
+      "loss": 0.3804,
+      "step": 412928
+    },
+    {
+      "epoch": 5.076220663982235,
+      "grad_norm": 13.829341888427734,
+      "learning_rate": 1.4923792323947125e-05,
+      "loss": 0.4603,
+      "step": 413184
+    },
+    {
+      "epoch": 5.0793658106584845,
+      "grad_norm": 0.9439744353294373,
+      "learning_rate": 1.4920647206250923e-05,
+      "loss": 0.3926,
+      "step": 413440
+    },
+    {
+      "epoch": 5.082510957334734,
+      "grad_norm": 15.285249710083008,
+      "learning_rate": 1.4917502088554721e-05,
+      "loss": 0.4066,
+      "step": 413696
+    },
+    {
+      "epoch": 5.085656104010983,
+      "grad_norm": 140.6249542236328,
+      "learning_rate": 1.4914356970858519e-05,
+      "loss": 0.3287,
+      "step": 413952
+    },
+    {
+      "epoch": 5.088801250687233,
+      "grad_norm": 0.36269086599349976,
+      "learning_rate": 1.4911211853162319e-05,
+      "loss": 0.4096,
+      "step": 414208
+    },
+    {
+      "epoch": 5.091946397363483,
+      "grad_norm": 1.0556248426437378,
+      "learning_rate": 1.4908066735466119e-05,
+      "loss": 0.3463,
+      "step": 414464
+    },
+    {
+      "epoch": 5.095091544039732,
+      "grad_norm": 64.4525146484375,
+      "learning_rate": 1.4904921617769917e-05,
+      "loss": 0.3485,
+      "step": 414720
+    },
+    {
+      "epoch": 5.098236690715981,
+      "grad_norm": 0.11732959747314453,
+      "learning_rate": 1.4901776500073715e-05,
+      "loss": 0.3436,
+      "step": 414976
+    },
+    {
+      "epoch": 5.101381837392231,
+      "grad_norm": 0.2561857998371124,
+      "learning_rate": 1.4898631382377513e-05,
+      "loss": 0.4747,
+      "step": 415232
+    },
+    {
+      "epoch": 5.104526984068481,
+      "grad_norm": 55.87223434448242,
+      "learning_rate": 1.4895486264681311e-05,
+      "loss": 0.3505,
+      "step": 415488
+    },
+    {
+      "epoch": 5.107672130744731,
+      "grad_norm": 0.025031782686710358,
+      "learning_rate": 1.4892341146985113e-05,
+      "loss": 0.4242,
+      "step": 415744
+    },
+    {
+      "epoch": 5.1108172774209795,
+      "grad_norm": 0.727285623550415,
+      "learning_rate": 1.488919602928891e-05,
+      "loss": 0.3706,
+      "step": 416000
+    },
+    {
+      "epoch": 5.113962424097229,
+      "grad_norm": 0.1500883847475052,
+      "learning_rate": 1.4886050911592709e-05,
+      "loss": 0.3627,
+      "step": 416256
+    },
+    {
+      "epoch": 5.117107570773479,
+      "grad_norm": 8.881159782409668,
+      "learning_rate": 1.4882905793896507e-05,
+      "loss": 0.3393,
+      "step": 416512
+    },
+    {
+      "epoch": 5.120252717449729,
+      "grad_norm": 1.1535494327545166,
+      "learning_rate": 1.4879760676200305e-05,
+      "loss": 0.3572,
+      "step": 416768
+    },
+    {
+      "epoch": 5.123397864125978,
+      "grad_norm": 0.4272177815437317,
+      "learning_rate": 1.4876615558504103e-05,
+      "loss": 0.3183,
+      "step": 417024
+    },
+    {
+      "epoch": 5.126543010802227,
+      "grad_norm": 22.624832153320312,
+      "learning_rate": 1.4873470440807903e-05,
+      "loss": 0.3261,
+      "step": 417280
+    },
+    {
+      "epoch": 5.129688157478477,
+      "grad_norm": 0.2767268717288971,
+      "learning_rate": 1.4870325323111703e-05,
+      "loss": 0.3896,
+      "step": 417536
+    },
+    {
+      "epoch": 5.132833304154726,
+      "grad_norm": 0.32466772198677063,
+      "learning_rate": 1.48671802054155e-05,
+      "loss": 0.4055,
+      "step": 417792
+    },
+    {
+      "epoch": 5.135978450830976,
+      "grad_norm": 0.024814318865537643,
+      "learning_rate": 1.4864035087719299e-05,
+      "loss": 0.396,
+      "step": 418048
+    },
+    {
+      "epoch": 5.139123597507226,
+      "grad_norm": 245.53958129882812,
+      "learning_rate": 1.4860889970023099e-05,
+      "loss": 0.3996,
+      "step": 418304
+    },
+    {
+      "epoch": 5.142268744183475,
+      "grad_norm": 117.71217346191406,
+      "learning_rate": 1.4857744852326897e-05,
+      "loss": 0.4319,
+      "step": 418560
+    },
+    {
+      "epoch": 5.145413890859724,
+      "grad_norm": 116.89668273925781,
+      "learning_rate": 1.4854599734630695e-05,
+      "loss": 0.3192,
+      "step": 418816
+    },
+    {
+      "epoch": 5.148559037535974,
+      "grad_norm": 0.3126812279224396,
+      "learning_rate": 1.4851454616934495e-05,
+      "loss": 0.3639,
+      "step": 419072
+    },
+    {
+      "epoch": 5.151704184212224,
+      "grad_norm": 21.277118682861328,
+      "learning_rate": 1.4848309499238293e-05,
+      "loss": 0.4359,
+      "step": 419328
+    },
+    {
+      "epoch": 5.1548493308884735,
+      "grad_norm": 94.19546508789062,
+      "learning_rate": 1.4845164381542093e-05,
+      "loss": 0.4494,
+      "step": 419584
+    },
+    {
+      "epoch": 5.157994477564722,
+      "grad_norm": 25.620561599731445,
+      "learning_rate": 1.484201926384589e-05,
+      "loss": 0.3887,
+      "step": 419840
+    },
+    {
+      "epoch": 5.161139624240972,
+      "grad_norm": 12.36883544921875,
+      "learning_rate": 1.4838874146149689e-05,
+      "loss": 0.3452,
+      "step": 420096
+    },
+    {
+      "epoch": 5.164284770917222,
+      "grad_norm": 208.77377319335938,
+      "learning_rate": 1.4835729028453487e-05,
+      "loss": 0.3444,
+      "step": 420352
+    },
+    {
+      "epoch": 5.167429917593472,
+      "grad_norm": 178.72891235351562,
+      "learning_rate": 1.4832583910757285e-05,
+      "loss": 0.3766,
+      "step": 420608
+    },
+    {
+      "epoch": 5.170575064269721,
+      "grad_norm": 4.52494478225708,
+      "learning_rate": 1.4829438793061087e-05,
+      "loss": 0.3614,
+      "step": 420864
+    },
+    {
+      "epoch": 5.17372021094597,
+      "grad_norm": 0.02147882618010044,
+      "learning_rate": 1.4826293675364885e-05,
+      "loss": 0.4036,
+      "step": 421120
+    },
+    {
+      "epoch": 5.17686535762222,
+      "grad_norm": 13.29345703125,
+      "learning_rate": 1.4823148557668683e-05,
+      "loss": 0.3316,
+      "step": 421376
+    },
+    {
+      "epoch": 5.18001050429847,
+      "grad_norm": 9.457071304321289,
+      "learning_rate": 1.4820003439972481e-05,
+      "loss": 0.3508,
+      "step": 421632
+    },
+    {
+      "epoch": 5.183155650974719,
+      "grad_norm": 0.3050714135169983,
+      "learning_rate": 1.4816858322276279e-05,
+      "loss": 0.4215,
+      "step": 421888
+    },
+    {
+      "epoch": 5.1863007976509685,
+      "grad_norm": 53.147090911865234,
+      "learning_rate": 1.4813713204580079e-05,
+      "loss": 0.3824,
+      "step": 422144
+    },
+    {
+      "epoch": 5.189445944327218,
+      "grad_norm": 2.5130209922790527,
+      "learning_rate": 1.4810568086883879e-05,
+      "loss": 0.4298,
+      "step": 422400
+    },
+    {
+      "epoch": 5.192591091003468,
+      "grad_norm": 12.828927993774414,
+      "learning_rate": 1.4807422969187677e-05,
+      "loss": 0.3611,
+      "step": 422656
+    },
+    {
+      "epoch": 5.195736237679717,
+      "grad_norm": 29.501089096069336,
+      "learning_rate": 1.4804277851491475e-05,
+      "loss": 0.3302,
+      "step": 422912
+    },
+    {
+      "epoch": 5.198881384355967,
+      "grad_norm": 36.43281555175781,
+      "learning_rate": 1.4801132733795273e-05,
+      "loss": 0.4196,
+      "step": 423168
+    },
+    {
+      "epoch": 5.2020265310322165,
+      "grad_norm": 0.019974226132035255,
+      "learning_rate": 1.4797987616099073e-05,
+      "loss": 0.3614,
+      "step": 423424
+    },
+    {
+      "epoch": 5.205171677708465,
+      "grad_norm": 187.4714813232422,
+      "learning_rate": 1.4794842498402871e-05,
+      "loss": 0.4227,
+      "step": 423680
+    },
+    {
+      "epoch": 5.208316824384715,
+      "grad_norm": 62.586612701416016,
+      "learning_rate": 1.4791697380706669e-05,
+      "loss": 0.3552,
+      "step": 423936
+    },
+    {
+      "epoch": 5.211461971060965,
+      "grad_norm": 0.13307207822799683,
+      "learning_rate": 1.4788552263010469e-05,
+      "loss": 0.2686,
+      "step": 424192
+    },
+    {
+      "epoch": 5.214607117737215,
+      "grad_norm": 58.886043548583984,
+      "learning_rate": 1.4785407145314269e-05,
+      "loss": 0.3572,
+      "step": 424448
+    },
+    {
+      "epoch": 5.2177522644134635,
+      "grad_norm": 38.93104934692383,
+      "learning_rate": 1.4782262027618067e-05,
+      "loss": 0.4399,
+      "step": 424704
+    },
+    {
+      "epoch": 5.220897411089713,
+      "grad_norm": 0.048439305275678635,
+      "learning_rate": 1.4779116909921865e-05,
+      "loss": 0.4742,
+      "step": 424960
+    },
+    {
+      "epoch": 5.224042557765963,
+      "grad_norm": 243.83067321777344,
+      "learning_rate": 1.4775971792225663e-05,
+      "loss": 0.4012,
+      "step": 425216
+    },
+    {
+      "epoch": 5.227187704442213,
+      "grad_norm": 0.05511689558625221,
+      "learning_rate": 1.4772826674529461e-05,
+      "loss": 0.3506,
+      "step": 425472
+    },
+    {
+      "epoch": 5.230332851118462,
+      "grad_norm": 57.89175033569336,
+      "learning_rate": 1.4769681556833263e-05,
+      "loss": 0.3984,
+      "step": 425728
+    },
+    {
+      "epoch": 5.2334779977947115,
+      "grad_norm": 70.34017181396484,
+      "learning_rate": 1.476653643913706e-05,
+      "loss": 0.3375,
+      "step": 425984
+    },
+    {
+      "epoch": 5.236623144470961,
+      "grad_norm": 128.1634521484375,
+      "learning_rate": 1.4763391321440859e-05,
+      "loss": 0.4836,
+      "step": 426240
+    },
+    {
+      "epoch": 5.239768291147211,
+      "grad_norm": 29.05970001220703,
+      "learning_rate": 1.4760246203744657e-05,
+      "loss": 0.358,
+      "step": 426496
+    },
+    {
+      "epoch": 5.24291343782346,
+      "grad_norm": 132.31077575683594,
+      "learning_rate": 1.4757101086048455e-05,
+      "loss": 0.3388,
+      "step": 426752
+    },
+    {
+      "epoch": 5.24605858449971,
+      "grad_norm": 0.2453315705060959,
+      "learning_rate": 1.4753955968352253e-05,
+      "loss": 0.4024,
+      "step": 427008
+    },
+    {
+      "epoch": 5.249203731175959,
+      "grad_norm": 3.2180874347686768,
+      "learning_rate": 1.4750810850656053e-05,
+      "loss": 0.3768,
+      "step": 427264
+    },
+    {
+      "epoch": 5.252348877852209,
+      "grad_norm": 130.93499755859375,
+      "learning_rate": 1.4747665732959853e-05,
+      "loss": 0.3487,
+      "step": 427520
+    },
+    {
+      "epoch": 5.255494024528458,
+      "grad_norm": 1.1053277254104614,
+      "learning_rate": 1.474452061526365e-05,
+      "loss": 0.4751,
+      "step": 427776
+    },
+    {
+      "epoch": 5.258639171204708,
+      "grad_norm": 4.604104518890381,
+      "learning_rate": 1.4741375497567449e-05,
+      "loss": 0.3788,
+      "step": 428032
+    },
+    {
+      "epoch": 5.261784317880958,
+      "grad_norm": 3.67272686958313,
+      "learning_rate": 1.4738230379871247e-05,
+      "loss": 0.3562,
+      "step": 428288
+    },
+    {
+      "epoch": 5.264929464557207,
+      "grad_norm": 53.248435974121094,
+      "learning_rate": 1.4735085262175047e-05,
+      "loss": 0.3696,
+      "step": 428544
+    },
+    {
+      "epoch": 5.268074611233456,
+      "grad_norm": 67.01683807373047,
+      "learning_rate": 1.4731940144478845e-05,
+      "loss": 0.3929,
+      "step": 428800
+    },
+    {
+      "epoch": 5.271219757909706,
+      "grad_norm": 47.91697692871094,
+      "learning_rate": 1.4728795026782645e-05,
+      "loss": 0.4941,
+      "step": 429056
+    },
+    {
+      "epoch": 5.274364904585956,
+      "grad_norm": 22.96300506591797,
+      "learning_rate": 1.4725649909086443e-05,
+      "loss": 0.3532,
+      "step": 429312
+    },
+    {
+      "epoch": 5.2775100512622055,
+      "grad_norm": 2.7295775413513184,
+      "learning_rate": 1.4722504791390243e-05,
+      "loss": 0.3443,
+      "step": 429568
+    },
+    {
+      "epoch": 5.280655197938454,
+      "grad_norm": 47.19913864135742,
+      "learning_rate": 1.471935967369404e-05,
+      "loss": 0.391,
+      "step": 429824
+    },
+    {
+      "epoch": 5.283800344614704,
+      "grad_norm": 0.045273683965206146,
+      "learning_rate": 1.4716214555997839e-05,
+      "loss": 0.4448,
+      "step": 430080
+    },
+    {
+      "epoch": 5.286945491290954,
+      "grad_norm": 17.242761611938477,
+      "learning_rate": 1.4713069438301637e-05,
+      "loss": 0.3505,
+      "step": 430336
+    },
+    {
+      "epoch": 5.290090637967204,
+      "grad_norm": 0.4532771408557892,
+      "learning_rate": 1.4709924320605435e-05,
+      "loss": 0.4833,
+      "step": 430592
+    },
+    {
+      "epoch": 5.293235784643453,
+      "grad_norm": 46.612953186035156,
+      "learning_rate": 1.4706779202909237e-05,
+      "loss": 0.3405,
+      "step": 430848
+    },
+    {
+      "epoch": 5.296380931319702,
+      "grad_norm": 17.724462509155273,
+      "learning_rate": 1.4703634085213035e-05,
+      "loss": 0.3697,
+      "step": 431104
+    },
+    {
+      "epoch": 5.299526077995952,
+      "grad_norm": 0.061910126358270645,
+      "learning_rate": 1.4700488967516833e-05,
+      "loss": 0.3496,
+      "step": 431360
+    },
+    {
+      "epoch": 5.302671224672201,
+      "grad_norm": 0.15961693227291107,
+      "learning_rate": 1.4697343849820631e-05,
+      "loss": 0.3556,
+      "step": 431616
+    },
+    {
+      "epoch": 5.305816371348451,
+      "grad_norm": 0.24715426564216614,
+      "learning_rate": 1.4694198732124429e-05,
+      "loss": 0.3967,
+      "step": 431872
+    },
+    {
+      "epoch": 5.3089615180247005,
+      "grad_norm": 79.31776428222656,
+      "learning_rate": 1.4691053614428227e-05,
+      "loss": 0.3607,
+      "step": 432128
+    },
+    {
+      "epoch": 5.31210666470095,
+      "grad_norm": 0.9846069812774658,
+      "learning_rate": 1.4687908496732029e-05,
+      "loss": 0.3595,
+      "step": 432384
+    },
+    {
+      "epoch": 5.315251811377199,
+      "grad_norm": 13.298562049865723,
+      "learning_rate": 1.4684763379035827e-05,
+      "loss": 0.4267,
+      "step": 432640
+    },
+    {
+      "epoch": 5.318396958053449,
+      "grad_norm": 0.03253224492073059,
+      "learning_rate": 1.4681618261339625e-05,
+      "loss": 0.3795,
+      "step": 432896
+    },
+    {
+      "epoch": 5.321542104729699,
+      "grad_norm": 20.27620506286621,
+      "learning_rate": 1.4678473143643423e-05,
+      "loss": 0.4036,
+      "step": 433152
+    },
+    {
+      "epoch": 5.3246872514059485,
+      "grad_norm": 168.25357055664062,
+      "learning_rate": 1.4675328025947221e-05,
+      "loss": 0.3189,
+      "step": 433408
+    },
+    {
+      "epoch": 5.327832398082197,
+      "grad_norm": 1.046290397644043,
+      "learning_rate": 1.4672182908251021e-05,
+      "loss": 0.3847,
+      "step": 433664
+    },
+    {
+      "epoch": 5.330977544758447,
+      "grad_norm": 32.317588806152344,
+      "learning_rate": 1.4669037790554819e-05,
+      "loss": 0.4664,
+      "step": 433920
+    },
+    {
+      "epoch": 5.334122691434697,
+      "grad_norm": 119.66520690917969,
+      "learning_rate": 1.4665892672858619e-05,
+      "loss": 0.3981,
+      "step": 434176
+    },
+    {
+      "epoch": 5.337267838110947,
+      "grad_norm": 0.05644058436155319,
+      "learning_rate": 1.4662747555162417e-05,
+      "loss": 0.3928,
+      "step": 434432
+    },
+    {
+      "epoch": 5.3404129847871955,
+      "grad_norm": 334.09564208984375,
+      "learning_rate": 1.4659602437466217e-05,
+      "loss": 0.4249,
+      "step": 434688
+    },
+    {
+      "epoch": 5.343558131463445,
+      "grad_norm": 2.991253137588501,
+      "learning_rate": 1.4656457319770015e-05,
+      "loss": 0.4011,
+      "step": 434944
+    },
+    {
+      "epoch": 5.346703278139695,
+      "grad_norm": 14.594536781311035,
+      "learning_rate": 1.4653312202073813e-05,
+      "loss": 0.3848,
+      "step": 435200
+    },
+    {
+      "epoch": 5.349848424815945,
+      "grad_norm": 15.55915355682373,
+      "learning_rate": 1.4650167084377611e-05,
+      "loss": 0.368,
+      "step": 435456
+    },
+    {
+      "epoch": 5.352993571492194,
+      "grad_norm": 33.24541473388672,
+      "learning_rate": 1.464702196668141e-05,
+      "loss": 0.415,
+      "step": 435712
+    },
+    {
+      "epoch": 5.3561387181684434,
+      "grad_norm": 87.2530517578125,
+      "learning_rate": 1.464387684898521e-05,
+      "loss": 0.4209,
+      "step": 435968
+    },
+    {
+      "epoch": 5.359283864844693,
+      "grad_norm": 0.04254389926791191,
+      "learning_rate": 1.4640731731289009e-05,
+      "loss": 0.4503,
+      "step": 436224
+    },
+    {
+      "epoch": 5.362429011520943,
+      "grad_norm": 0.004545422736555338,
+      "learning_rate": 1.4637586613592807e-05,
+      "loss": 0.3627,
+      "step": 436480
+    },
+    {
+      "epoch": 5.365574158197192,
+      "grad_norm": 28.76519775390625,
+      "learning_rate": 1.4634441495896605e-05,
+      "loss": 0.5301,
+      "step": 436736
+    },
+    {
+      "epoch": 5.368719304873442,
+      "grad_norm": 0.06790630519390106,
+      "learning_rate": 1.4631296378200403e-05,
+      "loss": 0.3046,
+      "step": 436992
+    },
+    {
+      "epoch": 5.371864451549691,
+      "grad_norm": 515.1839599609375,
+      "learning_rate": 1.4628151260504201e-05,
+      "loss": 0.3755,
+      "step": 437248
+    },
+    {
+      "epoch": 5.37500959822594,
+      "grad_norm": 2.8838603496551514,
+      "learning_rate": 1.4625006142808003e-05,
+      "loss": 0.3255,
+      "step": 437504
+    },
+    {
+      "epoch": 5.37815474490219,
+      "grad_norm": 6.0704240798950195,
+      "learning_rate": 1.46218610251118e-05,
+      "loss": 0.4298,
+      "step": 437760
+    },
+    {
+      "epoch": 5.38129989157844,
+      "grad_norm": 0.13719062507152557,
+      "learning_rate": 1.4618715907415599e-05,
+      "loss": 0.3559,
+      "step": 438016
+    },
+    {
+      "epoch": 5.38444503825469,
+      "grad_norm": 14.808082580566406,
+      "learning_rate": 1.4615570789719397e-05,
+      "loss": 0.452,
+      "step": 438272
+    },
+    {
+      "epoch": 5.387590184930938,
+      "grad_norm": 0.07143834978342056,
+      "learning_rate": 1.4612425672023197e-05,
+      "loss": 0.4016,
+      "step": 438528
+    },
+    {
+      "epoch": 5.390735331607188,
+      "grad_norm": 18.020097732543945,
+      "learning_rate": 1.4609280554326995e-05,
+      "loss": 0.3946,
+      "step": 438784
+    },
+    {
+      "epoch": 5.393880478283438,
+      "grad_norm": 0.3657359480857849,
+      "learning_rate": 1.4606135436630793e-05,
+      "loss": 0.3895,
+      "step": 439040
+    },
+    {
+      "epoch": 5.397025624959688,
+      "grad_norm": 142.16893005371094,
+      "learning_rate": 1.4602990318934593e-05,
+      "loss": 0.4017,
+      "step": 439296
+    },
+    {
+      "epoch": 5.400170771635937,
+      "grad_norm": 325.0826721191406,
+      "learning_rate": 1.4599845201238391e-05,
+      "loss": 0.4252,
+      "step": 439552
+    },
+    {
+      "epoch": 5.403315918312186,
+      "grad_norm": 135.35899353027344,
+      "learning_rate": 1.459670008354219e-05,
+      "loss": 0.4463,
+      "step": 439808
+    },
+    {
+      "epoch": 5.406461064988436,
+      "grad_norm": 0.2491970658302307,
+      "learning_rate": 1.4593554965845989e-05,
+      "loss": 0.3808,
+      "step": 440064
+    },
+    {
+      "epoch": 5.409606211664686,
+      "grad_norm": 12.762040138244629,
+      "learning_rate": 1.4590409848149787e-05,
+      "loss": 0.3927,
+      "step": 440320
+    },
+    {
+      "epoch": 5.412751358340935,
+      "grad_norm": 0.42027613520622253,
+      "learning_rate": 1.4587264730453585e-05,
+      "loss": 0.3585,
+      "step": 440576
+    },
+    {
+      "epoch": 5.415896505017185,
+      "grad_norm": 120.40250396728516,
+      "learning_rate": 1.4584119612757387e-05,
+      "loss": 0.4268,
+      "step": 440832
+    },
+    {
+      "epoch": 5.419041651693434,
+      "grad_norm": 0.3726077973842621,
+      "learning_rate": 1.4580974495061185e-05,
+      "loss": 0.4139,
+      "step": 441088
+    },
+    {
+      "epoch": 5.422186798369684,
+      "grad_norm": 1.6762504577636719,
+      "learning_rate": 1.4577829377364983e-05,
+      "loss": 0.3128,
+      "step": 441344
+    },
+    {
+      "epoch": 5.425331945045933,
+      "grad_norm": 2.037259101867676,
+      "learning_rate": 1.4574684259668781e-05,
+      "loss": 0.348,
+      "step": 441600
+    },
+    {
+      "epoch": 5.428477091722183,
+      "grad_norm": 0.2665565311908722,
+      "learning_rate": 1.4571539141972579e-05,
+      "loss": 0.3637,
+      "step": 441856
+    },
+    {
+      "epoch": 5.4316222383984325,
+      "grad_norm": 0.05366134271025658,
+      "learning_rate": 1.4568394024276377e-05,
+      "loss": 0.4117,
+      "step": 442112
+    },
+    {
+      "epoch": 5.434767385074682,
+      "grad_norm": 10.613019943237305,
+      "learning_rate": 1.4565248906580175e-05,
+      "loss": 0.4109,
+      "step": 442368
+    },
+    {
+      "epoch": 5.437912531750931,
+      "grad_norm": 45.673370361328125,
+      "learning_rate": 1.4562103788883977e-05,
+      "loss": 0.3548,
+      "step": 442624
+    },
+    {
+      "epoch": 5.441057678427181,
+      "grad_norm": 0.41616290807724,
+      "learning_rate": 1.4558958671187775e-05,
+      "loss": 0.4213,
+      "step": 442880
+    },
+    {
+      "epoch": 5.444202825103431,
+      "grad_norm": 147.30850219726562,
+      "learning_rate": 1.4555813553491573e-05,
+      "loss": 0.3721,
+      "step": 443136
+    },
+    {
+      "epoch": 5.4473479717796796,
+      "grad_norm": 49.96714401245117,
+      "learning_rate": 1.4552668435795371e-05,
+      "loss": 0.349,
+      "step": 443392
+    },
+    {
+      "epoch": 5.450493118455929,
+      "grad_norm": 40.10234069824219,
+      "learning_rate": 1.4549523318099171e-05,
+      "loss": 0.4013,
+      "step": 443648
+    },
+    {
+      "epoch": 5.453638265132179,
+      "grad_norm": 23.951541900634766,
+      "learning_rate": 1.4546378200402969e-05,
+      "loss": 0.4042,
+      "step": 443904
+    },
+    {
+      "epoch": 5.456783411808429,
+      "grad_norm": 0.06460706144571304,
+      "learning_rate": 1.4543233082706769e-05,
+      "loss": 0.3543,
+      "step": 444160
+    },
+    {
+      "epoch": 5.459928558484679,
+      "grad_norm": 59.21656799316406,
+      "learning_rate": 1.4540087965010567e-05,
+      "loss": 0.4205,
+      "step": 444416
+    },
+    {
+      "epoch": 5.4630737051609275,
+      "grad_norm": 0.7488933801651001,
+      "learning_rate": 1.4536942847314365e-05,
+      "loss": 0.4082,
+      "step": 444672
+    },
+    {
+      "epoch": 5.466218851837177,
+      "grad_norm": 0.13989563286304474,
+      "learning_rate": 1.4533797729618165e-05,
+      "loss": 0.3482,
+      "step": 444928
+    },
+    {
+      "epoch": 5.469363998513427,
+      "grad_norm": 20.191158294677734,
+      "learning_rate": 1.4530652611921963e-05,
+      "loss": 0.4626,
+      "step": 445184
+    },
+    {
+      "epoch": 5.472509145189676,
+      "grad_norm": 0.03903035447001457,
+      "learning_rate": 1.4527507494225761e-05,
+      "loss": 0.4527,
+      "step": 445440
+    },
+    {
+      "epoch": 5.475654291865926,
+      "grad_norm": 2.1459619998931885,
+      "learning_rate": 1.452436237652956e-05,
+      "loss": 0.353,
+      "step": 445696
+    },
+    {
+      "epoch": 5.478799438542175,
+      "grad_norm": 0.0663018450140953,
+      "learning_rate": 1.452121725883336e-05,
+      "loss": 0.3979,
+      "step": 445952
+    },
+    {
+      "epoch": 5.481944585218425,
+      "grad_norm": 9.152742385864258,
+      "learning_rate": 1.4518072141137159e-05,
+      "loss": 0.4389,
+      "step": 446208
+    },
+    {
+      "epoch": 5.485089731894674,
+      "grad_norm": 13.338661193847656,
+      "learning_rate": 1.4514927023440957e-05,
+      "loss": 0.3815,
+      "step": 446464
+    },
+    {
+      "epoch": 5.488234878570924,
+      "grad_norm": 0.7131269574165344,
+      "learning_rate": 1.4511781905744755e-05,
+      "loss": 0.3519,
+      "step": 446720
+    },
+    {
+      "epoch": 5.491380025247174,
+      "grad_norm": 22.403228759765625,
+      "learning_rate": 1.4508636788048553e-05,
+      "loss": 0.3772,
+      "step": 446976
+    },
+    {
+      "epoch": 5.494525171923423,
+      "grad_norm": 103.68403625488281,
+      "learning_rate": 1.4505491670352351e-05,
+      "loss": 0.3783,
+      "step": 447232
+    },
+    {
+      "epoch": 5.497670318599672,
+      "grad_norm": 0.04808255657553673,
+      "learning_rate": 1.4502346552656153e-05,
+      "loss": 0.4817,
+      "step": 447488
+    },
+    {
+      "epoch": 5.500815465275922,
+      "grad_norm": 0.030783984810113907,
+      "learning_rate": 1.449920143495995e-05,
+      "loss": 0.3647,
+      "step": 447744
+    },
+    {
+      "epoch": 5.503960611952172,
+      "grad_norm": 4.150572299957275,
+      "learning_rate": 1.4496056317263749e-05,
+      "loss": 0.4314,
+      "step": 448000
+    },
+    {
+      "epoch": 5.5071057586284216,
+      "grad_norm": 1281.9638671875,
+      "learning_rate": 1.4492911199567547e-05,
+      "loss": 0.4489,
+      "step": 448256
+    },
+    {
+      "epoch": 5.51025090530467,
+      "grad_norm": 0.27607831358909607,
+      "learning_rate": 1.4489766081871345e-05,
+      "loss": 0.3715,
+      "step": 448512
+    },
+    {
+      "epoch": 5.51339605198092,
+      "grad_norm": 0.3387153744697571,
+      "learning_rate": 1.4486620964175145e-05,
+      "loss": 0.4245,
+      "step": 448768
+    },
+    {
+      "epoch": 5.51654119865717,
+      "grad_norm": 9.522908210754395,
+      "learning_rate": 1.4483475846478943e-05,
+      "loss": 0.3468,
+      "step": 449024
+    },
+    {
+      "epoch": 5.519686345333419,
+      "grad_norm": 0.7314357757568359,
+      "learning_rate": 1.4480330728782743e-05,
+      "loss": 0.3193,
+      "step": 449280
+    },
+    {
+      "epoch": 5.522831492009669,
+      "grad_norm": 17.625904083251953,
+      "learning_rate": 1.4477185611086541e-05,
+      "loss": 0.3994,
+      "step": 449536
+    },
+    {
+      "epoch": 5.525976638685918,
+      "grad_norm": 0.5833144187927246,
+      "learning_rate": 1.447404049339034e-05,
+      "loss": 0.385,
+      "step": 449792
+    },
+    {
+      "epoch": 5.529121785362168,
+      "grad_norm": 0.17006878554821014,
+      "learning_rate": 1.4470895375694139e-05,
+      "loss": 0.374,
+      "step": 450048
+    },
+    {
+      "epoch": 5.532266932038418,
+      "grad_norm": 4.848082065582275,
+      "learning_rate": 1.4467750257997937e-05,
+      "loss": 0.3889,
+      "step": 450304
+    },
+    {
+      "epoch": 5.535412078714667,
+      "grad_norm": 75.98141479492188,
+      "learning_rate": 1.4464605140301735e-05,
+      "loss": 0.3742,
+      "step": 450560
+    },
+    {
+      "epoch": 5.5385572253909165,
+      "grad_norm": 67.52642059326172,
+      "learning_rate": 1.4461460022605535e-05,
+      "loss": 0.3944,
+      "step": 450816
+    },
+    {
+      "epoch": 5.541702372067166,
+      "grad_norm": 0.7687580585479736,
+      "learning_rate": 1.4458314904909335e-05,
+      "loss": 0.4032,
+      "step": 451072
+    },
+    {
+      "epoch": 5.544847518743415,
+      "grad_norm": 0.22134076058864594,
+      "learning_rate": 1.4455169787213133e-05,
+      "loss": 0.4084,
+      "step": 451328
+    },
+    {
+      "epoch": 5.547992665419665,
+      "grad_norm": 0.9399652481079102,
+      "learning_rate": 1.4452024669516931e-05,
+      "loss": 0.4152,
+      "step": 451584
+    },
+    {
+      "epoch": 5.551137812095915,
+      "grad_norm": 10.746188163757324,
+      "learning_rate": 1.4448879551820729e-05,
+      "loss": 0.3555,
+      "step": 451840
+    },
+    {
+      "epoch": 5.5542829587721645,
+      "grad_norm": 0.3087367117404938,
+      "learning_rate": 1.4445734434124527e-05,
+      "loss": 0.4091,
+      "step": 452096
+    },
+    {
+      "epoch": 5.557428105448414,
+      "grad_norm": 39.65850067138672,
+      "learning_rate": 1.4442589316428325e-05,
+      "loss": 0.4756,
+      "step": 452352
+    },
+    {
+      "epoch": 5.560573252124663,
+      "grad_norm": 8.794588088989258,
+      "learning_rate": 1.4439444198732127e-05,
+      "loss": 0.438,
+      "step": 452608
+    },
+    {
+      "epoch": 5.563718398800913,
+      "grad_norm": 97.39041137695312,
+      "learning_rate": 1.4436299081035925e-05,
+      "loss": 0.4218,
+      "step": 452864
+    },
+    {
+      "epoch": 5.566863545477163,
+      "grad_norm": 385.8796691894531,
+      "learning_rate": 1.4433153963339723e-05,
+      "loss": 0.3366,
+      "step": 453120
+    },
+    {
+      "epoch": 5.5700086921534115,
+      "grad_norm": 0.43009307980537415,
+      "learning_rate": 1.4430008845643521e-05,
+      "loss": 0.326,
+      "step": 453376
+    },
+    {
+      "epoch": 5.573153838829661,
+      "grad_norm": 48.30652618408203,
+      "learning_rate": 1.442686372794732e-05,
+      "loss": 0.2944,
+      "step": 453632
+    },
+    {
+      "epoch": 5.576298985505911,
+      "grad_norm": 108.22708129882812,
+      "learning_rate": 1.4423718610251119e-05,
+      "loss": 0.4329,
+      "step": 453888
+    },
+    {
+      "epoch": 5.579444132182161,
+      "grad_norm": 73.95587158203125,
+      "learning_rate": 1.4420573492554917e-05,
+      "loss": 0.3716,
+      "step": 454144
+    },
+    {
+      "epoch": 5.58258927885841,
+      "grad_norm": 0.036468349397182465,
+      "learning_rate": 1.4417428374858717e-05,
+      "loss": 0.4341,
+      "step": 454400
+    },
+    {
+      "epoch": 5.5857344255346595,
+      "grad_norm": 97.75952911376953,
+      "learning_rate": 1.4414283257162515e-05,
+      "loss": 0.3951,
+      "step": 454656
+    },
+    {
+      "epoch": 5.588879572210909,
+      "grad_norm": 94.74980926513672,
+      "learning_rate": 1.4411138139466315e-05,
+      "loss": 0.3899,
+      "step": 454912
+    },
+    {
+      "epoch": 5.592024718887159,
+      "grad_norm": 44.84490966796875,
+      "learning_rate": 1.4407993021770113e-05,
+      "loss": 0.3535,
+      "step": 455168
+    },
+    {
+      "epoch": 5.595169865563408,
+      "grad_norm": 19.707807540893555,
+      "learning_rate": 1.4404847904073911e-05,
+      "loss": 0.4532,
+      "step": 455424
+    },
+    {
+      "epoch": 5.598315012239658,
+      "grad_norm": 3.646214008331299,
+      "learning_rate": 1.440170278637771e-05,
+      "loss": 0.4631,
+      "step": 455680
+    },
+    {
+      "epoch": 5.601460158915907,
+      "grad_norm": 0.23348554968833923,
+      "learning_rate": 1.4398557668681509e-05,
+      "loss": 0.4022,
+      "step": 455936
+    },
+    {
+      "epoch": 5.604605305592157,
+      "grad_norm": 1.4638307094573975,
+      "learning_rate": 1.4395412550985309e-05,
+      "loss": 0.3879,
+      "step": 456192
+    },
+    {
+      "epoch": 5.607750452268406,
+      "grad_norm": 66.01060485839844,
+      "learning_rate": 1.4392267433289107e-05,
+      "loss": 0.3364,
+      "step": 456448
+    },
+    {
+      "epoch": 5.610895598944656,
+      "grad_norm": 4.419394016265869,
+      "learning_rate": 1.4389122315592905e-05,
+      "loss": 0.4214,
+      "step": 456704
+    },
+    {
+      "epoch": 5.614040745620906,
+      "grad_norm": 5.224905490875244,
+      "learning_rate": 1.4385977197896703e-05,
+      "loss": 0.3583,
+      "step": 456960
+    },
+    {
+      "epoch": 5.6171858922971545,
+      "grad_norm": 0.2397933155298233,
+      "learning_rate": 1.4382832080200501e-05,
+      "loss": 0.434,
+      "step": 457216
+    },
+    {
+      "epoch": 5.620331038973404,
+      "grad_norm": 0.3536325693130493,
+      "learning_rate": 1.43796869625043e-05,
+      "loss": 0.3072,
+      "step": 457472
+    },
+    {
+      "epoch": 5.623476185649654,
+      "grad_norm": 17.955657958984375,
+      "learning_rate": 1.4376541844808101e-05,
+      "loss": 0.3606,
+      "step": 457728
+    },
+    {
+      "epoch": 5.626621332325904,
+      "grad_norm": 1.9287688732147217,
+      "learning_rate": 1.4373396727111899e-05,
+      "loss": 0.3496,
+      "step": 457984
+    },
+    {
+      "epoch": 5.6297664790021535,
+      "grad_norm": 118.85681915283203,
+      "learning_rate": 1.4370251609415697e-05,
+      "loss": 0.3954,
+      "step": 458240
+    },
+    {
+      "epoch": 5.632911625678402,
+      "grad_norm": 0.06956590712070465,
+      "learning_rate": 1.4367106491719495e-05,
+      "loss": 0.3875,
+      "step": 458496
+    },
+    {
+      "epoch": 5.636056772354652,
+      "grad_norm": 1.9446299076080322,
+      "learning_rate": 1.4363961374023295e-05,
+      "loss": 0.3912,
+      "step": 458752
+    },
+    {
+      "epoch": 5.639201919030902,
+      "grad_norm": 3.3335297107696533,
+      "learning_rate": 1.4360816256327093e-05,
+      "loss": 0.4379,
+      "step": 459008
+    },
+    {
+      "epoch": 5.642347065707151,
+      "grad_norm": 38.55470275878906,
+      "learning_rate": 1.4357671138630893e-05,
+      "loss": 0.3427,
+      "step": 459264
+    },
+    {
+      "epoch": 5.645492212383401,
+      "grad_norm": 0.7164474725723267,
+      "learning_rate": 1.4354526020934691e-05,
+      "loss": 0.4199,
+      "step": 459520
+    },
+    {
+      "epoch": 5.64863735905965,
+      "grad_norm": 85.09379577636719,
+      "learning_rate": 1.4351380903238489e-05,
+      "loss": 0.4039,
+      "step": 459776
+    },
+    {
+      "epoch": 5.6517825057359,
+      "grad_norm": 102.90345001220703,
+      "learning_rate": 1.4348235785542289e-05,
+      "loss": 0.3981,
+      "step": 460032
+    },
+    {
+      "epoch": 5.654927652412149,
+      "grad_norm": 0.03342989832162857,
+      "learning_rate": 1.4345090667846087e-05,
+      "loss": 0.3576,
+      "step": 460288
+    },
+    {
+      "epoch": 5.658072799088399,
+      "grad_norm": 390.9976501464844,
+      "learning_rate": 1.4341945550149885e-05,
+      "loss": 0.4836,
+      "step": 460544
+    },
+    {
+      "epoch": 5.6612179457646485,
+      "grad_norm": 0.1759815812110901,
+      "learning_rate": 1.4338800432453683e-05,
+      "loss": 0.403,
+      "step": 460800
+    },
+    {
+      "epoch": 5.664363092440898,
+      "grad_norm": 1.1475474834442139,
+      "learning_rate": 1.4335655314757485e-05,
+      "loss": 0.4358,
+      "step": 461056
+    },
+    {
+      "epoch": 5.667508239117147,
+      "grad_norm": 0.40320727229118347,
+      "learning_rate": 1.4332510197061283e-05,
+      "loss": 0.3874,
+      "step": 461312
+    },
+    {
+      "epoch": 5.670653385793397,
+      "grad_norm": 135.46595764160156,
+      "learning_rate": 1.4329365079365081e-05,
+      "loss": 0.3745,
+      "step": 461568
+    },
+    {
+      "epoch": 5.673798532469647,
+      "grad_norm": 0.6997230648994446,
+      "learning_rate": 1.4326219961668879e-05,
+      "loss": 0.3955,
+      "step": 461824
+    },
+    {
+      "epoch": 5.6769436791458965,
+      "grad_norm": 0.44991162419319153,
+      "learning_rate": 1.4323074843972677e-05,
+      "loss": 0.3779,
+      "step": 462080
+    },
+    {
+      "epoch": 5.680088825822145,
+      "grad_norm": 1.1512393951416016,
+      "learning_rate": 1.4319929726276475e-05,
+      "loss": 0.3665,
+      "step": 462336
+    },
+    {
+      "epoch": 5.683233972498395,
+      "grad_norm": 0.20269031822681427,
+      "learning_rate": 1.4316784608580277e-05,
+      "loss": 0.4585,
+      "step": 462592
+    },
+    {
+      "epoch": 5.686379119174645,
+      "grad_norm": 114.74862670898438,
+      "learning_rate": 1.4313639490884075e-05,
+      "loss": 0.4321,
+      "step": 462848
+    },
+    {
+      "epoch": 5.689524265850894,
+      "grad_norm": 25.296741485595703,
+      "learning_rate": 1.4310494373187873e-05,
+      "loss": 0.4289,
+      "step": 463104
+    },
+    {
+      "epoch": 5.6926694125271435,
+      "grad_norm": 15.068272590637207,
+      "learning_rate": 1.4307349255491671e-05,
+      "loss": 0.4364,
+      "step": 463360
+    },
+    {
+      "epoch": 5.695814559203393,
+      "grad_norm": 230.223388671875,
+      "learning_rate": 1.430420413779547e-05,
+      "loss": 0.3434,
+      "step": 463616
+    },
+    {
+      "epoch": 5.698959705879643,
+      "grad_norm": 2.4799678325653076,
+      "learning_rate": 1.4301059020099269e-05,
+      "loss": 0.4089,
+      "step": 463872
+    },
+    {
+      "epoch": 5.702104852555893,
+      "grad_norm": 69.57925415039062,
+      "learning_rate": 1.4297913902403067e-05,
+      "loss": 0.4061,
+      "step": 464128
+    },
+    {
+      "epoch": 5.705249999232142,
+      "grad_norm": 15.480486869812012,
+      "learning_rate": 1.4294768784706867e-05,
+      "loss": 0.3296,
+      "step": 464384
+    },
+    {
+      "epoch": 5.7083951459083915,
+      "grad_norm": 481.4758605957031,
+      "learning_rate": 1.4291623667010665e-05,
+      "loss": 0.4244,
+      "step": 464640
+    },
+    {
+      "epoch": 5.711540292584641,
+      "grad_norm": 0.030323617160320282,
+      "learning_rate": 1.4288478549314463e-05,
+      "loss": 0.3304,
+      "step": 464896
+    },
+    {
+      "epoch": 5.71468543926089,
+      "grad_norm": 0.3148115873336792,
+      "learning_rate": 1.4285333431618263e-05,
+      "loss": 0.4335,
+      "step": 465152
+    },
+    {
+      "epoch": 5.71783058593714,
+      "grad_norm": 47.644927978515625,
+      "learning_rate": 1.4282188313922061e-05,
+      "loss": 0.3639,
+      "step": 465408
+    },
+    {
+      "epoch": 5.72097573261339,
+      "grad_norm": 54.57386779785156,
+      "learning_rate": 1.427904319622586e-05,
+      "loss": 0.4145,
+      "step": 465664
+    },
+    {
+      "epoch": 5.724120879289639,
+      "grad_norm": 0.9598469734191895,
+      "learning_rate": 1.4275898078529659e-05,
+      "loss": 0.417,
+      "step": 465920
+    },
+    {
+      "epoch": 5.727266025965888,
+      "grad_norm": 0.05439724773168564,
+      "learning_rate": 1.4272752960833459e-05,
+      "loss": 0.4067,
+      "step": 466176
+    },
+    {
+      "epoch": 5.730411172642138,
+      "grad_norm": 9.609537124633789,
+      "learning_rate": 1.4269607843137257e-05,
+      "loss": 0.3297,
+      "step": 466432
+    },
+    {
+      "epoch": 5.733556319318388,
+      "grad_norm": 31.04618263244629,
+      "learning_rate": 1.4266462725441055e-05,
+      "loss": 0.4084,
+      "step": 466688
+    },
+    {
+      "epoch": 5.736701465994638,
+      "grad_norm": 90.02703094482422,
+      "learning_rate": 1.4263317607744853e-05,
+      "loss": 0.4175,
+      "step": 466944
+    },
+    {
+      "epoch": 5.7398466126708865,
+      "grad_norm": 0.01569804921746254,
+      "learning_rate": 1.4260172490048651e-05,
+      "loss": 0.4606,
+      "step": 467200
+    },
+    {
+      "epoch": 5.742991759347136,
+      "grad_norm": 142.52288818359375,
+      "learning_rate": 1.425702737235245e-05,
+      "loss": 0.4713,
+      "step": 467456
+    },
+    {
+      "epoch": 5.746136906023386,
+      "grad_norm": 18.881750106811523,
+      "learning_rate": 1.4253882254656251e-05,
+      "loss": 0.4046,
+      "step": 467712
+    },
+    {
+      "epoch": 5.749282052699636,
+      "grad_norm": 493.8915100097656,
+      "learning_rate": 1.4250737136960049e-05,
+      "loss": 0.3566,
+      "step": 467968
+    },
+    {
+      "epoch": 5.752427199375885,
+      "grad_norm": 0.006592131219804287,
+      "learning_rate": 1.4247592019263847e-05,
+      "loss": 0.3106,
+      "step": 468224
+    },
+    {
+      "epoch": 5.755572346052134,
+      "grad_norm": 832.1047973632812,
+      "learning_rate": 1.4244446901567645e-05,
+      "loss": 0.4297,
+      "step": 468480
+    },
+    {
+      "epoch": 5.758717492728384,
+      "grad_norm": 142.4406280517578,
+      "learning_rate": 1.4241301783871443e-05,
+      "loss": 0.349,
+      "step": 468736
+    },
+    {
+      "epoch": 5.761862639404634,
+      "grad_norm": 0.8432340621948242,
+      "learning_rate": 1.4238156666175243e-05,
+      "loss": 0.3342,
+      "step": 468992
+    },
+    {
+      "epoch": 5.765007786080883,
+      "grad_norm": 14.462039947509766,
+      "learning_rate": 1.4235011548479043e-05,
+      "loss": 0.5093,
+      "step": 469248
+    },
+    {
+      "epoch": 5.768152932757133,
+      "grad_norm": 0.012599022127687931,
+      "learning_rate": 1.4231866430782841e-05,
+      "loss": 0.365,
+      "step": 469504
+    },
+    {
+      "epoch": 5.771298079433382,
+      "grad_norm": 67.51305389404297,
+      "learning_rate": 1.4228721313086639e-05,
+      "loss": 0.3652,
+      "step": 469760
+    },
+    {
+      "epoch": 5.774443226109632,
+      "grad_norm": 24.07037925720215,
+      "learning_rate": 1.4225576195390439e-05,
+      "loss": 0.4156,
+      "step": 470016
+    },
+    {
+      "epoch": 5.777588372785881,
+      "grad_norm": 18.413991928100586,
+      "learning_rate": 1.4222431077694237e-05,
+      "loss": 0.4378,
+      "step": 470272
+    },
+    {
+      "epoch": 5.780733519462131,
+      "grad_norm": 53.740211486816406,
+      "learning_rate": 1.4219285959998035e-05,
+      "loss": 0.4627,
+      "step": 470528
+    },
+    {
+      "epoch": 5.7838786661383805,
+      "grad_norm": 83.43620300292969,
+      "learning_rate": 1.4216140842301833e-05,
+      "loss": 0.3027,
+      "step": 470784
+    },
+    {
+      "epoch": 5.787023812814629,
+      "grad_norm": 0.8465031385421753,
+      "learning_rate": 1.4212995724605633e-05,
+      "loss": 0.3817,
+      "step": 471040
+    },
+    {
+      "epoch": 5.790168959490879,
+      "grad_norm": 131.7412872314453,
+      "learning_rate": 1.4209850606909433e-05,
+      "loss": 0.4122,
+      "step": 471296
+    },
+    {
+      "epoch": 5.793314106167129,
+      "grad_norm": 0.09065686911344528,
+      "learning_rate": 1.4206705489213231e-05,
+      "loss": 0.3812,
+      "step": 471552
+    },
+    {
+      "epoch": 5.796459252843379,
+      "grad_norm": 3.338660717010498,
+      "learning_rate": 1.4203560371517029e-05,
+      "loss": 0.3984,
+      "step": 471808
+    },
+    {
+      "epoch": 5.7996043995196285,
+      "grad_norm": 49.132503509521484,
+      "learning_rate": 1.4200415253820827e-05,
+      "loss": 0.3528,
+      "step": 472064
+    },
+    {
+      "epoch": 5.802749546195877,
+      "grad_norm": 115.07254028320312,
+      "learning_rate": 1.4197270136124625e-05,
+      "loss": 0.4797,
+      "step": 472320
+    },
+    {
+      "epoch": 5.805894692872127,
+      "grad_norm": 0.0621442086994648,
+      "learning_rate": 1.4194125018428427e-05,
+      "loss": 0.4183,
+      "step": 472576
+    },
+    {
+      "epoch": 5.809039839548377,
+      "grad_norm": 277.4070739746094,
+      "learning_rate": 1.4190979900732225e-05,
+      "loss": 0.3779,
+      "step": 472832
+    },
+    {
+      "epoch": 5.812184986224626,
+      "grad_norm": 69.28643798828125,
+      "learning_rate": 1.4187834783036023e-05,
+      "loss": 0.4028,
+      "step": 473088
+    },
+    {
+      "epoch": 5.8153301329008755,
+      "grad_norm": 152.26426696777344,
+      "learning_rate": 1.4184689665339821e-05,
+      "loss": 0.4223,
+      "step": 473344
+    },
+    {
+      "epoch": 5.818475279577125,
+      "grad_norm": 7.379275321960449,
+      "learning_rate": 1.418154454764362e-05,
+      "loss": 0.3621,
+      "step": 473600
+    },
+    {
+      "epoch": 5.821620426253375,
+      "grad_norm": 69.53936767578125,
+      "learning_rate": 1.4178399429947417e-05,
+      "loss": 0.4222,
+      "step": 473856
+    },
+    {
+      "epoch": 5.824765572929624,
+      "grad_norm": 0.8465156555175781,
+      "learning_rate": 1.4175254312251217e-05,
+      "loss": 0.4324,
+      "step": 474112
+    },
+    {
+      "epoch": 5.827910719605874,
+      "grad_norm": 258.32391357421875,
+      "learning_rate": 1.4172109194555017e-05,
+      "loss": 0.4017,
+      "step": 474368
+    },
+    {
+      "epoch": 5.8310558662821235,
+      "grad_norm": 183.59188842773438,
+      "learning_rate": 1.4168964076858815e-05,
+      "loss": 0.3786,
+      "step": 474624
+    },
+    {
+      "epoch": 5.834201012958373,
+      "grad_norm": 36.69376754760742,
+      "learning_rate": 1.4165818959162613e-05,
+      "loss": 0.376,
+      "step": 474880
+    },
+    {
+      "epoch": 5.837346159634622,
+      "grad_norm": 655.2278442382812,
+      "learning_rate": 1.4162673841466413e-05,
+      "loss": 0.4015,
+      "step": 475136
+    },
+    {
+      "epoch": 5.840491306310872,
+      "grad_norm": 0.08521991223096848,
+      "learning_rate": 1.4159528723770211e-05,
+      "loss": 0.3492,
+      "step": 475392
+    },
+    {
+      "epoch": 5.843636452987122,
+      "grad_norm": 0.2802114188671112,
+      "learning_rate": 1.415638360607401e-05,
+      "loss": 0.4448,
+      "step": 475648
+    },
+    {
+      "epoch": 5.846781599663371,
+      "grad_norm": 1.0108816623687744,
+      "learning_rate": 1.4153238488377807e-05,
+      "loss": 0.4117,
+      "step": 475904
+    },
+    {
+      "epoch": 5.84992674633962,
+      "grad_norm": 0.2940574884414673,
+      "learning_rate": 1.4150093370681607e-05,
+      "loss": 0.3197,
+      "step": 476160
+    },
+    {
+      "epoch": 5.85307189301587,
+      "grad_norm": 56.06767272949219,
+      "learning_rate": 1.4146948252985407e-05,
+      "loss": 0.3736,
+      "step": 476416
+    },
+    {
+      "epoch": 5.85621703969212,
+      "grad_norm": 101.5466079711914,
+      "learning_rate": 1.4143803135289205e-05,
+      "loss": 0.3967,
+      "step": 476672
+    },
+    {
+      "epoch": 5.859362186368369,
+      "grad_norm": 202.8722686767578,
+      "learning_rate": 1.4140658017593003e-05,
+      "loss": 0.4092,
+      "step": 476928
+    },
+    {
+      "epoch": 5.8625073330446185,
+      "grad_norm": 0.04954686015844345,
+      "learning_rate": 1.4137512899896801e-05,
+      "loss": 0.3926,
+      "step": 477184
+    },
+    {
+      "epoch": 5.865652479720868,
+      "grad_norm": 74.03327178955078,
+      "learning_rate": 1.41343677822006e-05,
+      "loss": 0.3803,
+      "step": 477440
+    },
+    {
+      "epoch": 5.868797626397118,
+      "grad_norm": 0.28958678245544434,
+      "learning_rate": 1.4131222664504401e-05,
+      "loss": 0.4394,
+      "step": 477696
+    },
+    {
+      "epoch": 5.871942773073368,
+      "grad_norm": 0.020548425614833832,
+      "learning_rate": 1.4128077546808199e-05,
+      "loss": 0.3713,
+      "step": 477952
+    },
+    {
+      "epoch": 5.875087919749617,
+      "grad_norm": 1.1551705598831177,
+      "learning_rate": 1.4124932429111997e-05,
+      "loss": 0.3889,
+      "step": 478208
+    },
+    {
+      "epoch": 5.878233066425866,
+      "grad_norm": 4.713179111480713,
+      "learning_rate": 1.4121787311415795e-05,
+      "loss": 0.4424,
+      "step": 478464
+    },
+    {
+      "epoch": 5.881378213102116,
+      "grad_norm": 33.48601150512695,
+      "learning_rate": 1.4118642193719593e-05,
+      "loss": 0.3696,
+      "step": 478720
+    },
+    {
+      "epoch": 5.884523359778365,
+      "grad_norm": 0.16305293142795563,
+      "learning_rate": 1.4115497076023391e-05,
+      "loss": 0.4266,
+      "step": 478976
+    },
+    {
+      "epoch": 5.887668506454615,
+      "grad_norm": 2.200655698776245,
+      "learning_rate": 1.4112351958327191e-05,
+      "loss": 0.3822,
+      "step": 479232
+    },
+    {
+      "epoch": 5.890813653130865,
+      "grad_norm": 0.7315693497657776,
+      "learning_rate": 1.4109206840630991e-05,
+      "loss": 0.3955,
+      "step": 479488
+    },
+    {
+      "epoch": 5.893958799807114,
+      "grad_norm": 82.9139404296875,
+      "learning_rate": 1.4106061722934789e-05,
+      "loss": 0.4516,
+      "step": 479744
+    },
+    {
+      "epoch": 5.897103946483363,
+      "grad_norm": 26.960159301757812,
+      "learning_rate": 1.4102916605238587e-05,
+      "loss": 0.4218,
+      "step": 480000
+    },
+    {
+      "epoch": 5.900249093159613,
+      "grad_norm": 0.41803106665611267,
+      "learning_rate": 1.4099771487542387e-05,
+      "loss": 0.5214,
+      "step": 480256
+    },
+    {
+      "epoch": 5.903394239835863,
+      "grad_norm": 293.9013366699219,
+      "learning_rate": 1.4096626369846185e-05,
+      "loss": 0.3473,
+      "step": 480512
+    },
+    {
+      "epoch": 5.9065393865121125,
+      "grad_norm": 172.8380584716797,
+      "learning_rate": 1.4093481252149983e-05,
+      "loss": 0.4249,
+      "step": 480768
+    },
+    {
+      "epoch": 5.909684533188361,
+      "grad_norm": 0.45075523853302,
+      "learning_rate": 1.4090336134453783e-05,
+      "loss": 0.3022,
+      "step": 481024
+    },
+    {
+      "epoch": 5.912829679864611,
+      "grad_norm": 12.004117012023926,
+      "learning_rate": 1.4087191016757583e-05,
+      "loss": 0.3877,
+      "step": 481280
+    },
+    {
+      "epoch": 5.915974826540861,
+      "grad_norm": 34.665771484375,
+      "learning_rate": 1.4084045899061381e-05,
+      "loss": 0.4662,
+      "step": 481536
+    },
+    {
+      "epoch": 5.919119973217111,
+      "grad_norm": 0.08787938952445984,
+      "learning_rate": 1.4080900781365179e-05,
+      "loss": 0.4703,
+      "step": 481792
+    },
+    {
+      "epoch": 5.92226511989336,
+      "grad_norm": 1.2439390420913696,
+      "learning_rate": 1.4077755663668977e-05,
+      "loss": 0.3925,
+      "step": 482048
+    },
+    {
+      "epoch": 5.925410266569609,
+      "grad_norm": 72.55826568603516,
+      "learning_rate": 1.4074610545972775e-05,
+      "loss": 0.3669,
+      "step": 482304
+    },
+    {
+      "epoch": 5.928555413245859,
+      "grad_norm": 32.59156036376953,
+      "learning_rate": 1.4071465428276573e-05,
+      "loss": 0.3072,
+      "step": 482560
+    },
+    {
+      "epoch": 5.931700559922108,
+      "grad_norm": 0.08439311385154724,
+      "learning_rate": 1.4068320310580375e-05,
+      "loss": 0.4249,
+      "step": 482816
+    },
+    {
+      "epoch": 5.934845706598358,
+      "grad_norm": 0.3488181233406067,
+      "learning_rate": 1.4065175192884173e-05,
+      "loss": 0.4258,
+      "step": 483072
+    },
+    {
+      "epoch": 5.9379908532746075,
+      "grad_norm": 86.25321197509766,
+      "learning_rate": 1.4062030075187971e-05,
+      "loss": 0.4663,
+      "step": 483328
+    },
+    {
+      "epoch": 5.941135999950857,
+      "grad_norm": 88.48367309570312,
+      "learning_rate": 1.405888495749177e-05,
+      "loss": 0.3511,
+      "step": 483584
+    },
+    {
+      "epoch": 5.944281146627107,
+      "grad_norm": 0.16091573238372803,
+      "learning_rate": 1.4055739839795567e-05,
+      "loss": 0.2617,
+      "step": 483840
+    },
+    {
+      "epoch": 5.947426293303356,
+      "grad_norm": 0.5184911489486694,
+      "learning_rate": 1.4052594722099367e-05,
+      "loss": 0.4471,
+      "step": 484096
+    },
+    {
+      "epoch": 5.950571439979606,
+      "grad_norm": 0.054081544280052185,
+      "learning_rate": 1.4049449604403167e-05,
+      "loss": 0.3492,
+      "step": 484352
+    },
+    {
+      "epoch": 5.9537165866558555,
+      "grad_norm": 1.2904483079910278,
+      "learning_rate": 1.4046304486706965e-05,
+      "loss": 0.4206,
+      "step": 484608
+    },
+    {
+      "epoch": 5.956861733332104,
+      "grad_norm": 13.672941207885742,
+      "learning_rate": 1.4043159369010763e-05,
+      "loss": 0.4187,
+      "step": 484864
+    },
+    {
+      "epoch": 5.960006880008354,
+      "grad_norm": 363.4638366699219,
+      "learning_rate": 1.4040014251314561e-05,
+      "loss": 0.4127,
+      "step": 485120
+    },
+    {
+      "epoch": 5.963152026684604,
+      "grad_norm": 0.9625645279884338,
+      "learning_rate": 1.4036869133618361e-05,
+      "loss": 0.4297,
+      "step": 485376
+    },
+    {
+      "epoch": 5.966297173360854,
+      "grad_norm": 0.040608711540699005,
+      "learning_rate": 1.403372401592216e-05,
+      "loss": 0.4077,
+      "step": 485632
+    },
+    {
+      "epoch": 5.969442320037103,
+      "grad_norm": 15.77828598022461,
+      "learning_rate": 1.4030578898225957e-05,
+      "loss": 0.3791,
+      "step": 485888
+    },
+    {
+      "epoch": 5.972587466713352,
+      "grad_norm": 0.1851249486207962,
+      "learning_rate": 1.4027433780529757e-05,
+      "loss": 0.3738,
+      "step": 486144
+    },
+    {
+      "epoch": 5.975732613389602,
+      "grad_norm": 0.5289295315742493,
+      "learning_rate": 1.4024288662833557e-05,
+      "loss": 0.4053,
+      "step": 486400
+    },
+    {
+      "epoch": 5.978877760065852,
+      "grad_norm": 0.6161154508590698,
+      "learning_rate": 1.4021143545137355e-05,
+      "loss": 0.3653,
+      "step": 486656
+    },
+    {
+      "epoch": 5.982022906742101,
+      "grad_norm": 67.921630859375,
+      "learning_rate": 1.4017998427441153e-05,
+      "loss": 0.3068,
+      "step": 486912
+    },
+    {
+      "epoch": 5.9851680534183505,
+      "grad_norm": 52.27008819580078,
+      "learning_rate": 1.4014853309744951e-05,
+      "loss": 0.4589,
+      "step": 487168
+    },
+    {
+      "epoch": 5.9883132000946,
+      "grad_norm": 0.03570324555039406,
+      "learning_rate": 1.401170819204875e-05,
+      "loss": 0.3869,
+      "step": 487424
+    },
+    {
+      "epoch": 5.99145834677085,
+      "grad_norm": 0.6653123497962952,
+      "learning_rate": 1.4008563074352551e-05,
+      "loss": 0.4658,
+      "step": 487680
+    },
+    {
+      "epoch": 5.994603493447099,
+      "grad_norm": 60.28636932373047,
+      "learning_rate": 1.4005417956656349e-05,
+      "loss": 0.3709,
+      "step": 487936
+    },
+    {
+      "epoch": 5.997748640123349,
+      "grad_norm": 545.3649291992188,
+      "learning_rate": 1.4002272838960147e-05,
+      "loss": 0.4226,
+      "step": 488192
+    },
+    {
+      "epoch": 6.0,
+      "eval_AGE": {
+        "f1": 0.9672360768551506,
+        "number": 32192,
+        "precision": 0.9484920871902061,
+        "recall": 0.9867358349900597
+      },
+      "eval_BUILDINGNUM": {
+        "f1": 0.9832519792331442,
+        "number": 28889,
+        "precision": 0.9798886138613861,
+        "recall": 0.9866385129287965
+      },
+      "eval_CITY": {
+        "f1": 0.9726022504018575,
+        "number": 41669,
+        "precision": 0.9651668399659703,
+        "recall": 0.9801531114257601
+      },
+      "eval_DATE": {
+        "f1": 0.997746454524687,
+        "number": 80940,
+        "precision": 0.9972108062744517,
+        "recall": 0.9982826785273041
+      },
+      "eval_EMAIL": {
+        "f1": 0.9962614462862082,
+        "number": 51445,
+        "precision": 0.9953818689847873,
+        "recall": 0.9971425794537856
+      },
+      "eval_GENDER": {
+        "f1": 0.974448482073879,
+        "number": 17457,
+        "precision": 0.973473587925909,
+        "recall": 0.9754253308128544
+      },
+      "eval_GIVENNAME": {
+        "f1": 0.9365210222588624,
+        "number": 100350,
+        "precision": 0.9334706161888168,
+        "recall": 0.9395914299950174
+      },
+      "eval_IDCARDNUM": {
+        "f1": 0.9049785748995008,
+        "number": 22068,
+        "precision": 0.8827889338964061,
+        "recall": 0.9283124886713794
+      },
+      "eval_PASSPORTNUM": {
+        "f1": 0.9242263349514563,
+        "number": 12959,
+        "precision": 0.9087180252069506,
+        "recall": 0.9402731692260206
+      },
+      "eval_SEX": {
+        "f1": 0.9694227341286166,
+        "number": 14570,
+        "precision": 0.9683604985618408,
+        "recall": 0.970487302676733
+      },
+      "eval_SOCIALNUM": {
+        "f1": 0.9280586914587484,
+        "number": 15327,
+        "precision": 0.9004663721158567,
+        "recall": 0.9573954459450643
+      },
+      "eval_STREET": {
+        "f1": 0.9810103469263542,
+        "number": 32803,
+        "precision": 0.9793116019078287,
+        "recall": 0.9827149955796726
+      },
+      "eval_SURNAME": {
+        "f1": 0.9306129607436204,
+        "number": 85192,
+        "precision": 0.926976695433423,
+        "recall": 0.9342778664663348
+      },
+      "eval_TELEPHONENUM": {
+        "f1": 0.9940678080618354,
+        "number": 36941,
+        "precision": 0.9924450920079866,
+        "recall": 0.9956958393113343
+      },
+      "eval_ZIPCODE": {
+        "f1": 0.9856799457604832,
+        "number": 24346,
+        "precision": 0.9860648661980516,
+        "recall": 0.9852953257208577
+      },
+      "eval_overall_accuracy": 0.9935988810850189,
+      "eval_overall_f1": 0.963716296062499,
+      "eval_overall_precision": 0.958534819652143,
+      "eval_overall_recall": 0.968954095132195,
+      "eval_runtime": 3717.4163,
+      "eval_samples_per_second": 87.583,
+      "eval_steps_per_second": 43.791,
+      "step": 488376
+    }
+  ],
+  "logging_steps": 256,
+  "max_steps": 1627920,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 20,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.3030150193655552e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}