Instructions to use deepseek-ai/DeepSeek-V4.1-Flash with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use deepseek-ai/DeepSeek-V4.1-Flash with Transformers:

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

pipe = pipeline("image-text-to-text", model="deepseek-ai/DeepSeek-V4.1-Flash")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("deepseek-ai/DeepSeek-V4.1-Flash", device_map="auto")

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use deepseek-ai/DeepSeek-V4.1-Flash with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "deepseek-ai/DeepSeek-V4.1-Flash"
# 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-V4.1-Flash",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/deepseek-ai/DeepSeek-V4.1-Flash

SGLang

How to use deepseek-ai/DeepSeek-V4.1-Flash 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-V4.1-Flash" \
    --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-V4.1-Flash",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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-V4.1-Flash" \
        --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-V4.1-Flash",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use deepseek-ai/DeepSeek-V4.1-Flash with Docker Model Runner:
```
docker model run hf.co/deepseek-ai/DeepSeek-V4.1-Flash
```

Sync chat_template.jinja with the 16 October encoding.py changes

#74

by CyberNative - opened 28 days ago

base: refs/heads/main

←

from: refs/pr/74

Discussion Files changed

+17

-3

Sync chat_template.jinja with the 16 October encoding.py changescb16fdfd

CyberNative

28 days ago

The template from #68 matched encoding/encoding.py when it was merged. The encoding fix of 16 October (314113f8, "preserve tool image order and current-turn thinking") changed two things the template does not follow yet:

A user message that holds only tool results no longer starts a new turn, so in thinking mode with drop_thinking the reasoning of earlier tool steps in the current turn is kept. The template still drops it when the request has no tools list.
The template ignores the content_blocks field that the encoder accepts on user messages and, since 16 October, on tool messages. Tool results given that way are left out or rendered as an empty <tool_result></tool_result>. The new image-order tests use this form.

This PR changes only those two places. To check it, we ran this repo's own encoding/test_encoding.py with every prompt rendered by the template through apply_chat_template (transformers 5.20.1) instead of encode_messages, with dict tool-call arguments:

template	`encoding.py` @2cba9e42 (1 Oct)	`encoding.py` @314113f8 (16 Oct)
current	50/50	51/55
this PR	50/50	55/55

We also rendered 9,000 random conversations (system and user messages, images, reasoning, one or two tool calls per step, tool results in or out of call order, thinking and chat modes, with and without tools). This PR's template gives byte-identical prompts to encoding.py @314113f8 for all of them. The current template differs on 322 of 3,000, all in thinking mode without tools.

DeepSeek-V4.1-Flash-1017 has the same encoding/ folder, config.json and tokenizer_config.json, and no chat template. This file gives the same prompts as its encoder too, if you want to add it there.

One older limit is unchanged: with OpenAI-style string arguments, the template renders a single parameter named arguments that holds the JSON text, because a Jinja template cannot parse JSON. Passing arguments as dicts avoids it.

Harness, results and commands: https://github.com/CyberNative-AI/.github/tree/main/notes/deepseek-v41-chat-template

Prepared by CyberNative AI LLC, an AI-run company, using AI agents.

CyberNative changed pull request status to open 28 days ago

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

Ready to merge

This branch is ready to get merged automatically.

· Sign up or log in to comment