Instructions to use MMOPD/Qwen3-4B-OT3-if with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MMOPD/Qwen3-4B-OT3-if with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="MMOPD/Qwen3-4B-OT3-if")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("MMOPD/Qwen3-4B-OT3-if")
model = AutoModelForCausalLM.from_pretrained("MMOPD/Qwen3-4B-OT3-if", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use MMOPD/Qwen3-4B-OT3-if with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MMOPD/Qwen3-4B-OT3-if"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MMOPD/Qwen3-4B-OT3-if",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/MMOPD/Qwen3-4B-OT3-if

SGLang

How to use MMOPD/Qwen3-4B-OT3-if with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "MMOPD/Qwen3-4B-OT3-if" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MMOPD/Qwen3-4B-OT3-if",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "MMOPD/Qwen3-4B-OT3-if" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MMOPD/Qwen3-4B-OT3-if",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use MMOPD/Qwen3-4B-OT3-if with Docker Model Runner:
```
docker model run hf.co/MMOPD/Qwen3-4B-OT3-if
```

Qwen3-4B-OT3-if

The Instruction Following domain teacher of No Pain, More Gain: Iterative Merging for Effective Multi-Teacher On-Policy Distillation (code).

Training


Initialization	Qwen3-4B-OT3 (Qwen3-4B-Base after SFT on OpenThoughts3-1.2M)
Method	GRPO
Data	`if` subset of MMOPD/domain-rl-data: 11,728 verifiable instruction-following prompts from the `rlvr1` blend of `nvidia/Nemotron-RL-Ultra-Training-Blends`
Recipe	200 steps, 64 prompts x 8 samples per step, 32,768-token rollouts at temperature 1.0, peak LR 3e-6, no KL penalty; reward = 1 if every instruction of the prompt is satisfied by the answer after the thinking block
Weights	bfloat16, cast from the fp32 training weights

Evaluation

IFBench prompt-level loose accuracy 59.4 (avg@3: three samples per prompt with seeds 42-44, temperature 1.0, top-p 1.0, 32,768-token completion cap).

Format

Qwen3 chat template in thinking mode. The reasoning follows the OpenThoughts3 format: the assistant turn starts with <think> (one space) and the reasoning, then \n</think>\n\n and the answer.

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "MMOPD/Qwen3-4B-OT3-if"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(repo, dtype="auto", device_map="auto")
messages = [{"role": "user", "content": "..."}]
text = tok.apply_chat_template(messages, tokenize=False, add_generation_prompt=True, enable_thinking=True)
out = model.generate(**tok(text, return_tensors="pt").to(model.device), max_new_tokens=32768,
                     do_sample=True, temperature=1.0, top_p=1.0)
print(tok.decode(out[0], skip_special_tokens=True))

Citation

@article{kim2026immopd,
  title   = {No Pain, More Gain: Iterative Merging for Effective Multi-Teacher On-Policy Distillation},
  author  = {Kim, Seonghyeon and Jang, Chaeyun and Lee, Noah and Kim, Boseop and Lee, Juho},
  journal = {arXiv preprint arXiv:2609.34745},
  year    = {2026}
}

Downloads last month: 72

Safetensors

Model size

4B params

Tensor type

BF16

Model tree for MMOPD/Qwen3-4B-OT3-if

Base model

Qwen/Qwen3-4B-Base

Finetuned

(545)

this model

Dataset used to train MMOPD/Qwen3-4B-OT3-if

Paper for MMOPD/Qwen3-4B-OT3-if

No Pain, More Gain: Iterative Merging for Effective Multi-Teacher On-Policy Distillation

Paper • 2609.34745 • Published Sep 28