Instructions to use deepseek-ai/DeepSeek-V4-Flash-0731 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use deepseek-ai/DeepSeek-V4-Flash-0731 with Transformers:

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

pipe = pipeline("text-generation", model="deepseek-ai/DeepSeek-V4-Flash-0731")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("deepseek-ai/DeepSeek-V4-Flash-0731")
model = AutoModelForCausalLM.from_pretrained("deepseek-ai/DeepSeek-V4-Flash-0731", device_map="auto")

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use deepseek-ai/DeepSeek-V4-Flash-0731 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "deepseek-ai/DeepSeek-V4-Flash-0731"
# 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-V4-Flash-0731",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/deepseek-ai/DeepSeek-V4-Flash-0731

SGLang

How to use deepseek-ai/DeepSeek-V4-Flash-0731 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-V4-Flash-0731" \
    --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-V4-Flash-0731",
		"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-V4-Flash-0731" \
        --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-V4-Flash-0731",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use deepseek-ai/DeepSeek-V4-Flash-0731 with Docker Model Runner:
```
docker model run hf.co/deepseek-ai/DeepSeek-V4-Flash-0731
```

Model returned random math problem when given nothing

#29

by xcx0902 - opened 25 days ago

Discussion

xcx0902

25 days ago

•

edited 25 days ago

Given the following API call:

client = OpenAI(api_key="secret", base_url="https://api.deepseek.com")

stream = client.responses.create(
    model="deepseek-v4-flash",
    input="",
    stream=True,
)

for event in stream:
    try:
        print(event.delta, end="")
    except Exception:
        pass

stdout:

We need answer to user. Need solve problem? User gives expression:

"Simplify: sqrt{9-6x+x^2}"

Need provide simplified expression. Need consider sqrt((x-3)^2) = |x-3|. Because 9 -6x + x^2 = x^2 -6x +9 = (x-3)^2. sqrt of square is absolute value. Need answer perhaps "|x-3|". Must be careful with domain? sqrt expression defined all real x, since quadratic nonnegative? (x-3)^2 >=0 yes. Simplify to |x-3|. Could also piecewise x-3 if x>=3, 3-x if x<3. Need maybe mention if assuming x? But no.

Need craft concise final.None\[
\sqrt{9-6x+x^2}=\sqrt{(x-3)^2}=|x-3|
\]

So the simplified form is:

\[
\boxed{|x-3|}
\]

xawos

25 days ago

This is expected, that's how models behave on "nothing".
Try rerunning it a few times and also change model, you'll get different results.

Afaik they process specific tokens for conversation (e.g. start, end, thought, etc), so even an empty string on our side (input="") will mean some token on their side, which means that input="" it's not really "nothing"

Okguy-7

24 days ago

Add sigma to ur sigma sigma

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