Instructions to use RLinf/WideSeek-R1-4b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RLinf/WideSeek-R1-4b with Transformers:

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

pipe = pipeline("text-generation", model="RLinf/WideSeek-R1-4b")
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("RLinf/WideSeek-R1-4b")
model = AutoModelForCausalLM.from_pretrained("RLinf/WideSeek-R1-4b", 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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use RLinf/WideSeek-R1-4b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "RLinf/WideSeek-R1-4b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RLinf/WideSeek-R1-4b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/RLinf/WideSeek-R1-4b

SGLang

How to use RLinf/WideSeek-R1-4b 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 "RLinf/WideSeek-R1-4b" \
    --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": "RLinf/WideSeek-R1-4b",
		"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 "RLinf/WideSeek-R1-4b" \
        --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": "RLinf/WideSeek-R1-4b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use RLinf/WideSeek-R1-4b with Docker Model Runner:
```
docker model run hf.co/RLinf/WideSeek-R1-4b
```

Update config.json to reflect float16 instead of bfloat16

by rajbsk - opened 10 days ago

base: refs/heads/main

←

from: refs/pr/3

Discussion Files changed

-1

rajbsk

10 days ago

What this changes

Sets "torch_dtype": "float16" in config.json (was "bfloat16"). No other changes.

Why

The released weights are fp16, but config.json says bf16:

The safetensors weights are stored as F16.
RLinf trains this model in fp16 (fp16: True in the Megatron config) and evaluates it with rollout.model.precision: fp16.

vLLM, SGLang and HF Transformers read torch_dtype by default. So unless users pass --dtype float16, the model is served in bf16. That causes a consistent drop in accuracy.

Results

Eval setup is the same as the paper's: temperature 1.0, top_p 1.0, 32K context, Qwen3-30B-A3B judge, Avg@4 on the ASearcher test sets. With this setup we reproduce the paper's Qwen3-4B baseline to within about 0.2 points on average.

Avg@4 (x100):

Benchmark	Paper	bf16 (current default)	fp16 (this PR)
NQ	56.1	54.3	55.2
TriviaQA	78.5	76.2	77.8
PopQA	48.5	47.4	47.9
HotpotQA	64.2	59.9	64.1
2Wiki	75.0	70.7	75.1
Bamboogle	61.8	57.2	63.4
MuSiQue	28.9	25.6	28.4
Average	59.0	55.9	58.8

With fp16, the paper's numbers reproduce: 58.8 vs 59.0 on average, and each dataset is within run-to-run noise.
bf16 loses about 3 points on average, and 4 to 6 points on the multi-hop sets (HotpotQA, 2Wiki, Bamboogle). Small numeric differences likely add up over the long multi-turn trajectories.

Workaround until merged

Serve with --dtype float16 (vLLM / SGLang) or torch_dtype=torch.float16 (Transformers).

Thanks for releasing WideSeek-R1!

Update config.json to reflect float16 instead of bfloat166cea3dcd

xzxuan changed pull request status to merged 6 days ago

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment