Instructions to use Yukang/Qwen2.5-3B-Open-R1-Code-GRPO with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Yukang/Qwen2.5-3B-Open-R1-Code-GRPO with Transformers:

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

pipe = pipeline("text-generation", model="Yukang/Qwen2.5-3B-Open-R1-Code-GRPO")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Yukang/Qwen2.5-3B-Open-R1-Code-GRPO")
model = AutoModelForCausalLM.from_pretrained("Yukang/Qwen2.5-3B-Open-R1-Code-GRPO")
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

vLLM

How to use Yukang/Qwen2.5-3B-Open-R1-Code-GRPO with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Yukang/Qwen2.5-3B-Open-R1-Code-GRPO"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Yukang/Qwen2.5-3B-Open-R1-Code-GRPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Yukang/Qwen2.5-3B-Open-R1-Code-GRPO

SGLang

How to use Yukang/Qwen2.5-3B-Open-R1-Code-GRPO 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 "Yukang/Qwen2.5-3B-Open-R1-Code-GRPO" \
    --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": "Yukang/Qwen2.5-3B-Open-R1-Code-GRPO",
		"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 "Yukang/Qwen2.5-3B-Open-R1-Code-GRPO" \
        --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": "Yukang/Qwen2.5-3B-Open-R1-Code-GRPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Yukang/Qwen2.5-3B-Open-R1-Code-GRPO with Docker Model Runner:
```
docker model run hf.co/Yukang/Qwen2.5-3B-Open-R1-Code-GRPO
```

Yukang commited on Jun 14, 2025

Commit

f2765ac

verified ·

1 Parent(s): a840137

Training in progress, step 510

Browse files

Files changed (8) hide show

README.md +1 -1
all_results.json +4 -4
config.json +1 -1
model-00001-of-00002.safetensors +1 -1
model-00002-of-00002.safetensors +1 -1
train_results.json +4 -4
trainer_state.json +0 -0
training_args.bin +1 -1

README.md CHANGED Viewed

@@ -29,7 +29,7 @@ print(output["generated_text"])
 ## Training procedure
-[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/chenyukang2020-nvidia/huggingface/runs/87jsozpo)
 This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).

 ## Training procedure
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/chenyukang2020-nvidia/huggingface/runs/3c0353cy)
 This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).

all_results.json CHANGED Viewed

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.054507823697492e-06,
-    "train_runtime": 74.8551,
     "train_samples": 35735,
-    "train_samples_per_second": 3419.94,
-    "train_steps_per_second": 6.68
 }

 {
     "total_flos": 0.0,
+    "train_loss": 9.052564064315237e-07,
+    "train_runtime": 75.296,
     "train_samples": 35735,
+    "train_samples_per_second": 3399.916,
+    "train_steps_per_second": 6.64
 }

config.json CHANGED Viewed

@@ -22,7 +22,7 @@
   "tie_word_embeddings": true,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.52.3",
-  "use_cache": true,
   "use_sliding_window": false,
   "vocab_size": 151936
 }

   "tie_word_embeddings": true,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.52.3",
+  "use_cache": false,
   "use_sliding_window": false,
   "vocab_size": 151936
 }

model-00001-of-00002.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:30a1c3e68897bce63060646578aae237a4e7fe4665fdbfe981a18e12955e3935
 size 4957560304

 version https://git-lfs.github.com/spec/v1
+oid sha256:6eb6df84914c787eed2bf2a200ef3c85bbe9d288a43a32a6b00caf7521fab0e1
 size 4957560304

model-00002-of-00002.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:97164a9d2e2aabf03e747f6ad265ad18373f5e2ec138503538a81b4441e37b8f
 size 1214366696

 version https://git-lfs.github.com/spec/v1
+oid sha256:307cd5439b2d80cfe9a640f0053a40ea50ff56c6b5574102626f485412b8f5d1
 size 1214366696

train_results.json CHANGED Viewed

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.054507823697492e-06,
-    "train_runtime": 74.8551,
     "train_samples": 35735,
-    "train_samples_per_second": 3419.94,
-    "train_steps_per_second": 6.68
 }

 {
     "total_flos": 0.0,
+    "train_loss": 9.052564064315237e-07,
+    "train_runtime": 75.296,
     "train_samples": 35735,
+    "train_samples_per_second": 3399.916,
+    "train_steps_per_second": 6.64
 }

trainer_state.json CHANGED Viewed

The diff for this file is too large to render. See raw diff

training_args.bin CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5da1eddcbb114c354e9593c8cba67202a1f2a74a6c5a77536d99af07d7db0751
 size 8504

 version https://git-lfs.github.com/spec/v1
+oid sha256:9846bdd09191086fb5e7169a23b49f5974ac89201b2747a86923aa998fb05bfd
 size 8504