Instructions to use Dogacel/Universal-DeepSeek-OCR-2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Dogacel/Universal-DeepSeek-OCR-2 with Transformers:

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

pipe = pipeline("image-text-to-text", model="Dogacel/Universal-DeepSeek-OCR-2", trust_remote_code=True)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("Dogacel/Universal-DeepSeek-OCR-2", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Dogacel/Universal-DeepSeek-OCR-2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Dogacel/Universal-DeepSeek-OCR-2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Dogacel/Universal-DeepSeek-OCR-2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Dogacel/Universal-DeepSeek-OCR-2

SGLang

How to use Dogacel/Universal-DeepSeek-OCR-2 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 "Dogacel/Universal-DeepSeek-OCR-2" \
    --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": "Dogacel/Universal-DeepSeek-OCR-2",
		"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 "Dogacel/Universal-DeepSeek-OCR-2" \
        --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": "Dogacel/Universal-DeepSeek-OCR-2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Dogacel/Universal-DeepSeek-OCR-2 with Docker Model Runner:
```
docker model run hf.co/Dogacel/Universal-DeepSeek-OCR-2
```

Universal DeepSeek-OCR 2 – CPU, MPS, CUDA Support

This repository uses the weights from the original DeepSeek-OCR 2 and modifies model to support inference on different devices such as CPU and MPS (Apple Metal GPU). By default runs on CPU.

Explore more human-like visual encoding.

Usage

Sample code available at: https://github.com/Dogacel/Universal-DeepSeek-OCR-2

mamba create -n deepseek-ocr-2 python=3.12.9
mamba activate deepseek-ocr-2

pip install torch==2.6.0  torchvision Pillow transformers==4.46.3 tokenizers==0.20.3 einops addict easydict

from transformers import AutoModel, AutoTokenizer
import torch

model_name = '.'

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModel.from_pretrained(model_name, trust_remote_code=True, use_safetensors=True)
model = model.eval().to("cpu").to(torch.float16)

# prompt = "<image>\nFree OCR. "
prompt = "<image>\n<|grounding|>Convert the document to markdown. "
image_file = 'samples/paper.png'
output_path = 'tmp'

res = model.infer(
    tokenizer, 
    prompt=prompt, 
    image_file=image_file, 
    output_path = output_path, 
    base_size = 1024, 
    image_size = 768, 
    crop_mode = True, 
    save_results = True,
    test_compress = True,
)

To change device type, you should update two things,

- model.eval().to("cpu").to(torch.float16)
+ model = model.eval().to("mps").to(torch.float16)

res = model.infer(
    tokenizer, 
    prompt=prompt, 
    image_file=image_file, 
    output_path = output_path, 
    base_size = 1024, 
    image_size = 768, 
    crop_mode = True, 
    save_results = True,
    test_compress = True,
+   device = "mps",
+   dtype = torch.float16,
)

For CUDA, you should also use bfloat16 to get as close as possible to the original implementation.

- model.eval().to("cpu").to(torch.float16)
+ model = model.eval().to("cuda").to(torch.bfloat16)

res = model.infer(
    tokenizer, 
    prompt=prompt, 
    image_file=image_file, 
    output_path = output_path, 
    base_size = 1024, 
    image_size = 768, 
    crop_mode = True, 
    save_results = True,
    test_compress = True,
+   device = "cuda",
+   dtype = torch.bfloat16,
)

Downloads last month: 1,155

Safetensors

Model size

3B params

Tensor type

BF16

Inference Providers NEW

Image-Text-to-Text

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Dogacel/Universal-DeepSeek-OCR-2

Base model

deepseek-ai/DeepSeek-OCR-2

Finetuned

(31)

this model