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

Libraries

How to use FlyCockpit/DeepSeek-V4-Flash-0731-vision with Transformers:

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

pipe = pipeline("image-text-to-text", model="FlyCockpit/DeepSeek-V4-Flash-0731-vision")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("FlyCockpit/DeepSeek-V4-Flash-0731-vision", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use FlyCockpit/DeepSeek-V4-Flash-0731-vision with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "FlyCockpit/DeepSeek-V4-Flash-0731-vision"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FlyCockpit/DeepSeek-V4-Flash-0731-vision",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/FlyCockpit/DeepSeek-V4-Flash-0731-vision

SGLang

How to use FlyCockpit/DeepSeek-V4-Flash-0731-vision 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 "FlyCockpit/DeepSeek-V4-Flash-0731-vision" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FlyCockpit/DeepSeek-V4-Flash-0731-vision",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "FlyCockpit/DeepSeek-V4-Flash-0731-vision" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FlyCockpit/DeepSeek-V4-Flash-0731-vision",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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

DeepSeek-V4-Flash-0731-vision / README.md

christopher-kapic

Fix HF markdown strikethrough from approximate ~ sizes

d8efc7d verified 5 days ago

preview code

Raw

History Blame Contribute Delete

3.32 kB

	---
	license: apache-2.0
	base_model: deepseek-ai/DeepSeek-V4-Flash-0731
	tags:
	- multimodal
	- vision-language
	- deepseek
	- deepseek-v4
	- screenshots
	- adapter
	- dgx-spark
	library_name: transformers
	pipeline_tag: image-text-to-text
	---

	# DeepSeek-V4-Flash — Vision Encoder

	<p align="center">
	<strong>Give your self-hosted DeepSeek eyes.</strong><br/>
	Official production <em>tower + projector</em> for
	<a href="https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731">DeepSeek-V4-Flash-0731</a>.
	</p>

	This repo is only the vision stack (about 900 MB total). The language model
	itself is still the upstream 0731 FP8 checkpoint (about 167 GB).

	```text
	image → DeepEncoderV2 tower → projector → splice at token 129279
	→ DeepSeek-V4-Flash-0731
	```

	Serve it on 2× DGX Spark:
	→ [FlyCockpit/DeepSeek-V4-Vision-2x-DGX-Sparks](https://github.com/FlyCockpit/DeepSeek-V4-Vision-2x-DGX-Sparks)

	---

	## Download

	```bash
	hf download FlyCockpit/DeepSeek-V4-Flash-0731-vision \
	--local-dir $HOME/dsv4-vision-assets
	```

	Or pull just the two weight files:

	```bash
	hf download FlyCockpit/DeepSeek-V4-Flash-0731-vision \
	adapter/latest.pt \
	tower/deepencoder_v2_tower.safetensors \
	--local-dir $HOME/dsv4-vision-assets
	```

	### Files

	\| path \| what \| size \| md5 \|
	\|---\|---\|---:\|---\|
	\| `adapter/latest.pt` \| use this — final projector \| 40 MB \| `d9b3b3bda8f790ecf7cd5a98e6fb93a5` \|
	\| `adapter/merged-004800-5af0c5.pt` \| same bytes (named by step) \| 40 MB \| `d9b3b3bda8f790ecf7cd5a98e6fb93a5` \|
	\| `tower/deepencoder_v2_tower.safetensors` \| frozen image tower \| 865 MB \| `2d5dba626d816cc367d28b32e744830e` \|
	\| `config.json` \| release metadata \| tiny \| — \|

	Final training step: 4800. Layout: tiles=2 (257 / 769 / 1281 image tokens).

	---

	## What it's for

	\| Strength \| Notes \|
	\|---\|---\|
	\| UI / screenshot understanding \| Primary target \|
	\| Documents & on-screen text \| Strong \|
	\| Everyday photos \| Decent \|
	\| Click agents on arbitrary real UIs \| Not claimed — see below \|

	Synthetic GUI coordinate probes pass; real-page computer-use does not
	transfer cleanly yet. Full write-up lives in the playbook's
	[capabilities doc](https://github.com/FlyCockpit/DeepSeek-V4-Vision-2x-DGX-Sparks/blob/main/docs/CAPABILITIES.md).

	---

	## Compatibility (important)

	This encoder maps into 0731's embedding space.

	\| base weights \| works? \|
	\|---\|---\|
	\| `deepseek-ai/DeepSeek-V4-Flash-0731` \| yes \|
	\| older Flash / DSpark / NVIDIA NVFP4 of the old line \| no \|

	Precision (FP8 vs a future 0731-matched quant) is fine. **Different training
	lineage is not.**

	---

	## Serve

	Use the open playbook (Mia-style, worker-first TP=2):

	https://github.com/FlyCockpit/DeepSeek-V4-Vision-2x-DGX-Sparks

	```bash
	git clone https://github.com/FlyCockpit/DeepSeek-V4-Vision-2x-DGX-Sparks.git
	cd DeepSeek-V4-Vision-2x-DGX-Sparks
	cp .env.vision.example .env.vision # set fabric IPs
	bash scripts/download-assets.sh
	bash scripts/build-image.sh
	bash scripts/start-vision.sh # run on head
	```

	API: `http://<head>:8899/v1` · model id: `deepseek-v4-flash-0731-vision`

	---

	## License

	- Adapter packaging in this repo: Apache-2.0
	- DeepEncoderV2 tower & DeepSeek-V4-Flash-0731 backbone: upstream terms apply