Normalize Motius Pipeline.from_pretrained metadata

README.md

@@ -1,5 +1,5 @@
 ---
-library_name: hftrainer
+library_name: motius
 pipeline_tag: other
 tags:
 - motion-generation
@@ -13,20 +13,20 @@ license: other
 
 # MoMask — Generative Masked Modeling of 3D Human Motions
 
-Text-to-motion baseline integrated into the hftrainer Model Zoo. Our
+Text-to-motion baseline integrated into the motius Model Zoo. Our
 reproduction is **fully self-contained and independent of `ref_repo`**: the
 RVQ-VAE tokenizer, the masked generative transformer, the residual transformer
 and the length estimator are all vendored into
-`hftrainer.models.motion.momask._momask`, preserving numerical parity with the
+`motius.models.motion.momask._momask`, preserving numerical parity with the
 released HumanML3D checkpoints. The CLIP ViT-B/32 text encoder is reloaded by
-name only for legacy lightweight artifacts; new hftrainer artifacts include
+name only for legacy lightweight artifacts; new motius artifacts include
 `clip.safetensors`.
 
 | | |
 |---|---|
 | **Task** | Text-to-Motion (T2M) |
 | **Bundle / Pipeline** | `MoMaskBundle` / `MoMaskPipeline` |
-| **Processed HF artifact** | [`ZeyuLing/hftrainer-momask-humanml3d`](https://huggingface.co/ZeyuLing/hftrainer-momask-humanml3d) |
+| **Processed HF artifact** | [`ZeyuLing/Motius-MoMask-HumanML3D`](https://huggingface.co/ZeyuLing/Motius-MoMask-HumanML3D) |
 | **Motion representation** | **HumanML3D-263** (263-dim, 20 fps, 22 joints) |
 | **Tokenizer** | RVQ-VAE, **6 residual quantizers**, codebook 512×512 |
 | **Generator** | MaskTransformer (masked iterative decoding) + ResidualTransformer |
@@ -38,20 +38,20 @@ name only for legacy lightweight artifacts; new hftrainer artifacts include
 
 ## Weights
 
-Self-contained hftrainer artifact (diffusers-style `from_pretrained`):
+Self-contained motius artifact (diffusers-style `from_pretrained`):
 
 | Artifact | Location | Contents | Status |
 |---|---|---|---|
-| MoMask HumanML3D | [`ZeyuLing/hftrainer-momask-humanml3d`](https://huggingface.co/ZeyuLing/hftrainer-momask-humanml3d) | `vq.safetensors` + `t2m_trans.safetensors` + `res_trans.safetensors` + `length_est.safetensors` + `clip.safetensors` + `momask_config.json` + `Mean.npy` / `Std.npy` | public Hub artifact |
+| MoMask HumanML3D | [`ZeyuLing/Motius-MoMask-HumanML3D`](https://huggingface.co/ZeyuLing/Motius-MoMask-HumanML3D) | `vq.safetensors` + `t2m_trans.safetensors` + `res_trans.safetensors` + `length_est.safetensors` + `clip.safetensors` + `momask_config.json` + `Mean.npy` / `Std.npy` | public Hub artifact |
 | local mirror | `checkpoints/momask/humanml3d` | same layout (produced by `convert_momask_checkpoint.py`, see below) | optional local cache |
 
 Use the published artifact directly from the Hub:
 
 ```python
-from hftrainer.pipelines.momask import MoMaskPipeline
+from motius.pipelines.momask import MoMaskPipeline
 
 pipe = MoMaskPipeline.from_pretrained(
-    "ZeyuLing/hftrainer-momask-humanml3d",
+    "ZeyuLing/Motius-MoMask-HumanML3D",
     device="cuda",
 )
 motions = pipe.infer_t2m(
@@ -74,7 +74,7 @@ python3 scripts/eval/convert_momask_checkpoint.py \
 **Use it:**
 
 ```python
-from hftrainer.pipelines.momask import MoMaskPipeline
+from motius.pipelines.momask import MoMaskPipeline
 
 pipe = MoMaskPipeline.from_pretrained("checkpoints/momask/humanml3d", device="cuda")
 # fixed length (frames @ 20 fps):
@@ -144,7 +144,7 @@ python3 scripts/eval/verify_evaluators.py --which hml263 \
 
 ### HumanML3D-263 evaluator (native space)
 
-| Metric | hftrainer | MoMask paper |
+| Metric | motius | MoMask paper |
 |---|---|---|
 | **FID** ↓ | **0.097** | 0.045 |
 | R-Precision Top-1 / 2 / 3 ↑ | **0.516 / 0.709 / 0.804** | 0.521 / 0.713 / 0.807 |
@@ -168,7 +168,7 @@ native HumanML3D-263 predictions are retargeted through the validated MDM-style
 chain: HML263 -> SMPL `motion_135` (IK refine-80, 20 -> 30 fps) ->
 MotionStreamer-272 -> `MotionStreamer272Evaluator`.
 
-| Metric | hftrainer | MS-272 GT/Real |
+| Metric | motius | MS-272 GT/Real |
 |---|---:|---:|
 | **FID** ↓ | **114.869** | 0.000 |
 | R-Precision Top-1 ↑ | **0.485** | 0.706 |
@@ -186,7 +186,7 @@ Run details: `n_repeats = 20`, `n_samples_used = 7392`,
 
 ## Implementation notes
 
-- **Vendored, ref_repo-independent**: `hftrainer/models/motion/momask/_momask/`
+- **Vendored, ref_repo-independent**: `motius/models/motion/momask/_momask/`
   holds the RVQ-VAE (`vq/`), the masked / residual transformers
   (`mask_transformer/`) and the masked iterative decoding entry point
   (`inference.py`). Imports are package-relative; training-only code paths are
@@ -200,3 +200,11 @@ Run details: `n_repeats = 20`, `n_samples_used = 7392`,
 - **Normalization travels with the checkpoint**: `Mean.npy` / `Std.npy` are the
   RVQ-VAE training stats, embedded in the artifact.
 - **Guidance**: classifier-free, base `cond_scale=4`, residual `cond_scale=5`.
+
+## Direct Loading
+
+```python
+from motius import Pipeline
+
+pipeline = Pipeline.from_pretrained("ZeyuLing/Motius-MoMask-HumanML3D")
+```

model_index.json

@@ -1,10 +1,10 @@
 {
   "_class_name": "MoMaskPipeline",
-  "_library_name": "hftrainer",
+  "_library_name": "motius",
   "model_type": "momask",
-  "format": "hftrainer-momask-v1",
-  "bundle_class": "hftrainer.models.motion.momask.bundle.MoMaskBundle",
-  "pipeline_class": "hftrainer.pipelines.momask.pipeline.MoMaskPipeline",
+  "format": "motius-momask-v1",
+  "bundle_class": "motius.models.momask.MoMaskBundle",
+  "pipeline_class": "motius.pipelines.momask.MoMaskPipeline",
   "artifacts": {
     "vq": "vq.safetensors",
     "t2m_transformer": "t2m_trans.safetensors",
@@ -29,7 +29,23 @@
     }
   },
   "api": {
-    "from_pretrained": "hftrainer.models.motion.momask.MoMaskBundle.from_pretrained",
-    "from_config": "hftrainer.models.motion.momask.MoMaskBundle.from_config"
-  }
-}
+    "loader": "motius.Pipeline.from_pretrained",
+    "task_methods": [
+      "infer_text_to_motion"
+    ]
+  },
+  "format_version": 1,
+  "tasks": [
+    "text_to_motion"
+  ],
+  "required_files": [
+    "Mean.npy",
+    "Std.npy",
+    "clip.safetensors",
+    "length_est.safetensors",
+    "momask_config.json",
+    "res_trans.safetensors",
+    "t2m_trans.safetensors",
+    "vq.safetensors"
+  ]
+}