Instructions to use meituan/DeepSeek-R1-Channel-INT8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use meituan/DeepSeek-R1-Channel-INT8 with Transformers:

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

pipe = pipeline("text-generation", model="meituan/DeepSeek-R1-Channel-INT8", 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("meituan/DeepSeek-R1-Channel-INT8", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("meituan/DeepSeek-R1-Channel-INT8", 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 meituan/DeepSeek-R1-Channel-INT8 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "meituan/DeepSeek-R1-Channel-INT8"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "meituan/DeepSeek-R1-Channel-INT8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/meituan/DeepSeek-R1-Channel-INT8

SGLang

How to use meituan/DeepSeek-R1-Channel-INT8 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 "meituan/DeepSeek-R1-Channel-INT8" \
    --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": "meituan/DeepSeek-R1-Channel-INT8",
		"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 "meituan/DeepSeek-R1-Channel-INT8" \
        --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": "meituan/DeepSeek-R1-Channel-INT8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use meituan/DeepSeek-R1-Channel-INT8 with Docker Model Runner:
```
docker model run hf.co/meituan/DeepSeek-R1-Channel-INT8
```

HandH1998 commited on Feb 27, 2025

Commit

95642c7

verified ·

1 Parent(s): c6bab3f

Update README.md

Browse files

Files changed (1) hide show

README.md +34 -0

README.md CHANGED Viewed

@@ -2,6 +2,40 @@
 license: mit
 library_name: transformers
 ---
 # DeepSeek-R1
 <!-- markdownlint-disable first-line-h1 -->
 <!-- markdownlint-disable html -->

 license: mit
 library_name: transformers
 ---
+# Channel-wise INT8 DeepSeek-R1
+The INT8 data type is both friendly and efficient for most hardware platforms.
+**We provide a channel-wise INT8 weight for DeepSeek-R1.**
+In benchmarking, we observe **no accuracy loss** and up to **50\%** performance enhancement.
+[SGLang](https://github.com/sgl-project/sglang/tree/main) will soon support the channel-wise INT8 quantization operation once our [PULL REQUEST](https://github.com/sgl-project/sglang/pull/3888) is merged.
+## 1. Benchmarking Result (detailed in [PULL REQUEST](https://github.com/sgl-project/sglang/pull/3888)):
+| Model  | Config | Accuracy (GSM8K) | Accuracy (MMLU) | Output Throughput(qps=128) |
+|--------|--------|-------------------|----------------|------------------------------|
+| BF16 R1 | A100\*32  | 95.5              | 87.1           | 3342.29                       |
+| INT8 R1 | (A100\*16)x2 | **95.6**              | **87.2**           | **5035.82 (+50%)**                |
+## 2. Quantization Process
+We apply INT8 quantization to the BF16 checkpoints.
+The quantization scales are determined by dividing the channnel-wise maximum of element values by the INT8 type maximum.
+To generate this weight, run the provided script in the ``./inference`` directory:
+``
+python3 bf16_cast_channel_int8.py --input-bf16-hf-path /path/to/bf16-weights/ --output-int8-hf-path /path/to/save-int8-weight/
+``
+## 3. Trouble Shooting
+Before inference, you should confirm that there is no attribute "quantization_config" in `config.json`.
+---
 # DeepSeek-R1
 <!-- markdownlint-disable first-line-h1 -->
 <!-- markdownlint-disable html -->