Instructions to use wanzhenchn/DeepSeek-R1-AWQ with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use wanzhenchn/DeepSeek-R1-AWQ with Transformers:

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

pipe = pipeline("text-generation", model="wanzhenchn/DeepSeek-R1-AWQ", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("wanzhenchn/DeepSeek-R1-AWQ", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("wanzhenchn/DeepSeek-R1-AWQ", trust_remote_code=True)
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 wanzhenchn/DeepSeek-R1-AWQ with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/wanzhenchn/DeepSeek-R1-AWQ

SGLang

How to use wanzhenchn/DeepSeek-R1-AWQ 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 "wanzhenchn/DeepSeek-R1-AWQ" \
    --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": "wanzhenchn/DeepSeek-R1-AWQ",
		"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 "wanzhenchn/DeepSeek-R1-AWQ" \
        --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": "wanzhenchn/DeepSeek-R1-AWQ",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use wanzhenchn/DeepSeek-R1-AWQ with Docker Model Runner:
```
docker model run hf.co/wanzhenchn/DeepSeek-R1-AWQ
```

Thank you

by davidsyoung - opened Apr 3, 2025

Discussion

davidsyoung

Apr 3, 2025

I wanted to say thank you, as the group size of 128 on this quant allows me to run it on 16x3090 with 16k ctx, which is awesome.

Did you ever test 256?

wanzhenchn

Owner Apr 3, 2025

Thanks for your feedback. I have not conveted and tested the version with group size of 256, you can convert it step by step: DeepSeek-R1(fp8) -> BF16 -> AWQ(AutoAWQ, group_size=256)

Let me konw if you have any questions about it.

superahn

May 19, 2025

Can you tell me which GPUs and how many GPUs you used when performing AWQ quantization on DeepSeek-R1?

wanzhenchn

Owner May 25, 2025

Can you tell me which GPUs and how many GPUs you used when performing AWQ quantization on DeepSeek-R1?

8xB200 (180GB), you might want to try different types of GPUs.

wanzhenchn changed discussion status to closed May 25, 2025

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