deepseek-ai
/

DeepSeek-R1-0528

Text Generation

text-generation-inference

Model card Files Files and versions

Instructions to use deepseek-ai/DeepSeek-R1-0528 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use deepseek-ai/DeepSeek-R1-0528 with Transformers:

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

pipe = pipeline("text-generation", model="deepseek-ai/DeepSeek-R1-0528", 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("deepseek-ai/DeepSeek-R1-0528", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("deepseek-ai/DeepSeek-R1-0528", 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]:]))

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use deepseek-ai/DeepSeek-R1-0528 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "deepseek-ai/DeepSeek-R1-0528"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "deepseek-ai/DeepSeek-R1-0528",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/deepseek-ai/DeepSeek-R1-0528

How to use deepseek-ai/DeepSeek-R1-0528 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 "deepseek-ai/DeepSeek-R1-0528" \
    --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": "deepseek-ai/DeepSeek-R1-0528",
		"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 "deepseek-ai/DeepSeek-R1-0528" \
        --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": "deepseek-ai/DeepSeek-R1-0528",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use deepseek-ai/DeepSeek-R1-0528 with Docker Model Runner:
```
docker model run hf.co/deepseek-ai/DeepSeek-R1-0528
```

Resources

View closed (10)

Add EvalEval community eval results

#117 opened 4 days ago by

rohit

#116 opened about 2 months ago by

Install & run deepseek-ai/DeepSeek-R1-0528 easily using llmpm

#115 opened 3 months ago by

Add MMLU-Pro evaluation result (85.0)

#114 opened 4 months ago by

When I run on A800 , it throws error that ValueError: FP8 quantized models is only supported on GPUs with compute capability >= 8.9 (e.g 4090/H100), actual = 8.0

#113 opened 7 months ago by

BTC TRADE ETMEK

#112 opened 7 months ago by

Distill to Qwen/Qwen3-30B-A3B

#111 opened 9 months ago by

Request: DOI

#110 opened 10 months ago by

Update README.md

#109 opened 10 months ago by

DeepSeek-R1-0528 Additional Distill Models

#108 opened 11 months ago by

Linux mint terminal.

#107 opened 11 months ago by

Make them stand together

#106 opened 11 months ago by

text gen langchain endpoint not working

#105 opened 11 months ago by

deepseek

#104 opened 11 months ago by

Voice identification

#102 opened 12 months ago by

With DeepSeek-R1-0528-Qwen3-8B only

#101 opened 12 months ago by

Feedback

#100 opened 12 months ago by

int4 quantization destorys function_call accuracy

#99 opened 12 months ago by

deepseek-r1

#98 opened 12 months ago by

Request: DOI

#96 opened 12 months ago by

comparing with Phi-4

#95 opened 12 months ago by

Teste agora

#94 opened 12 months ago by

Readme.md Update

#93 opened about 1 year ago by

Rename README.md to Left: left=140, top=250, right=170, bottom=280. Right: left=230, top=250, right=260, bottom=280.

#92 opened about 1 year ago by

Test

#91 opened about 1 year ago by

This discussion has been hidden

#90 opened about 1 year ago by deleted

Definitely an improvement, however, it spams Chinese letters in my code

#89 opened about 1 year ago by

I would like to ask, has the paper associated with this model been published?

#88 opened about 1 year ago by

A man in the court room with lawyers

#85 opened about 1 year ago by

Questions about deploying function calls locally

#84 opened about 1 year ago by

Any plans for 32B/70B distilled models?

#83 opened about 1 year ago by

https://huggingface.co/proxy/enzostvs-deepsite.hf.space/ not completing prompt task, any help?

#81 opened about 1 year ago by

Fix broken chat template

#80 opened about 1 year ago by

Qwen3 4B distill

#79 opened about 1 year ago by

Update tokenizer_config.json

#78 opened about 1 year ago by

Update README.md

#77 opened about 1 year ago by

does this model has tools support

#76 opened about 1 year ago by

刚部署满血deepseek r1 0528版本，推理性能提升这么多嘛？不是架构没变嘛？

#75 opened about 1 year ago by

期待V4，迫不及待了🤩

#74 opened about 1 year ago by

support function call?

#73 opened about 1 year ago by

This is by no means a minor upgrade.

#72 opened about 1 year ago by

后续DeepSeek更新发版，建议同步发布官方DeepSeek量化版（如AWQ、INT4等），进一步降低资源要求

#71 opened about 1 year ago by

New R1 is indeed stronger!

#70 opened about 1 year ago by

fixed

#69 opened about 1 year ago by

Do you have deepseek-r1-0528-awq plan?

#68 opened about 1 year ago by

README.md: Fixup model config!

#66 opened about 1 year ago by

DeepSeek-R1-0528 仍然使用 2024 年 12 月所发布的 DeepSeek V3 Base 模型作为基座

#65 opened about 1 year ago by

How to run 0528version on GPU which don't support FP8

#64 opened about 1 year ago by

There will be a 8b distilled model! cool

#63 opened about 1 year ago by

Create README.md

#62 opened about 1 year ago by