Instructions to use cmpatino/qwen-grpo-r4-s125 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cmpatino/qwen-grpo-r4-s125 with Transformers:

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

pipe = pipeline("text-generation", model="cmpatino/qwen-grpo-r4-s125")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("cmpatino/qwen-grpo-r4-s125")
model = AutoModelForCausalLM.from_pretrained("cmpatino/qwen-grpo-r4-s125", 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use cmpatino/qwen-grpo-r4-s125 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cmpatino/qwen-grpo-r4-s125"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cmpatino/qwen-grpo-r4-s125",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/cmpatino/qwen-grpo-r4-s125

SGLang

How to use cmpatino/qwen-grpo-r4-s125 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 "cmpatino/qwen-grpo-r4-s125" \
    --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": "cmpatino/qwen-grpo-r4-s125",
		"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 "cmpatino/qwen-grpo-r4-s125" \
        --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": "cmpatino/qwen-grpo-r4-s125",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use cmpatino/qwen-grpo-r4-s125 with Docker Model Runner:
```
docker model run hf.co/cmpatino/qwen-grpo-r4-s125
```

Qwen3-0.6B + GRPO on GSM8K -- intermediate checkpoint (step 125 of run r4)

inspect_evals/gsm8k, full 1319-sample test set, 10-shot, greedy: 0.7081 +/- 0.0125 (untrained Qwen/Qwen3-0.6B baseline: 0.4754 +/- 0.0138).

A statistical tie with the campaign's best model, cmpatino/qwen-grpo-r5 (0.7089 +/- 0.0125) -- see that model card for the full recipe, reward function, and caveats. Note this is the step-125 checkpoint, which scores higher than run r4's final step-129 weights (0.6907).

Runs in non-thinking mode: the chat template always emits an empty <think></think> block. Use the tokenizer shipped in this repo.

Downloads last month: 14

Safetensors

Model size

0.6B params

Tensor type

F32

Model tree for cmpatino/qwen-grpo-r4-s125

Base model

Qwen/Qwen3-0.6B-Base

Finetuned

Qwen/Qwen3-0.6B

Finetuned

(1253)

this model

cmpatino
/

qwen-grpo-r4-s125

Qwen3-0.6B + GRPO on GSM8K -- intermediate checkpoint (step 125 of run r4)

Model tree for cmpatino/qwen-grpo-r4-s125

Dataset used to train cmpatino/qwen-grpo-r4-s125