Instructions to use Qwen/QwQ-32B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Qwen/QwQ-32B with Transformers:

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

pipe = pipeline("text-generation", model="Qwen/QwQ-32B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Qwen/QwQ-32B")
model = AutoModelForCausalLM.from_pretrained("Qwen/QwQ-32B")
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]:]))

Inference
Local Apps Settings

vLLM

How to use Qwen/QwQ-32B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Qwen/QwQ-32B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Qwen/QwQ-32B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Qwen/QwQ-32B

SGLang

How to use Qwen/QwQ-32B 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 "Qwen/QwQ-32B" \
    --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": "Qwen/QwQ-32B",
		"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 "Qwen/QwQ-32B" \
        --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": "Qwen/QwQ-32B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Qwen/QwQ-32B with Docker Model Runner:
```
docker model run hf.co/Qwen/QwQ-32B
```

Too many "cross-validate" and "another method"

#43

by AaronFeng753 - opened Mar 7, 2025

Discussion

AaronFeng753

Mar 7, 2025

•

edited Mar 7, 2025

First, I want to express my admiration for the incredible capabilities of this model! It’s truly impressive how it approaches problems with such depth and self-reflection.

During a recent interaction:

Question: How many days are between 12-12-1971 and 18-4-2024? (answer 19121)

Chat history: https://pastebin.com/1CVd4CLP

I noticed that the model arrived at the correct answer on its first attempt but proceeded to "cross-validate" the solution multiple times (approximately 3-4 instances). While this iterative process is a key strength of reasoning models, in this case, the repeated validation seemed excessive. In one instance, the back-and-forth even led to a momentary detour toward an incorrect conclusion before course-correcting.

I completely understand that self-verification is part of the model’s design, but I wonder if there might be opportunities to refine this behavior in future iterations. For scenarios where the initial answer is already well-supported, reducing cross-validation to 1-2 concise checks might streamline the process while retaining the model’s core strengths.

I’m sharing this feedback in the spirit of continuous improvement—thank you for building such an amazing tool! I’d love to hear your thoughts on how the model’s reasoning balance could evolve.

melekuk

Mar 11, 2025

You may need to use system prompt to achieve that
https://www.reddit.com/r/LocalLLaMA/comments/1j4v3fi/prompts_for_qwq32b/

Yanngp

Mar 20, 2025

Same here : QwQ is often stuck in recursive reasoning loops, it took QwQ32B q8_0 7min to answer a logic that was handled correctly in just 44 sec by Phi4.

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