DeepSeek-OCR

Read our Guide How to: Run & Fine-tune DeepSeek-OCR!

<div> <p style="margin-top: 0;margin-bottom: 0;"> <em>This DeepSeek-OCR upload was edited to enable inference & fine-tuning on the latest transformers (no accuracy change). <a href="https://docs.unsloth.ai/new/deepseek-ocr-run-and-fine-tune#fine-tuning-deepseek-ocr">Read more</a></em> </p> <div style="display: flex; gap: 5px; align-items: center; "> <a href="https://github.com/unslothai/unsloth/"> <img src="https://github.com/unslothai/unsloth/raw/main/images/unsloth%20new%20logo.png" width="133"> </a> <a href="https://discord.gg/unsloth"> <img src="https://github.com/unslothai/unsloth/raw/main/images/Discord%20button.png" width="173"> </a> <a href="https://docs.unsloth.ai/new/deepseek-ocr"> <img src="https://raw.githubusercontent.com/unslothai/unsloth/refs/heads/main/images/documentation%20green%20button.png" width="143"> </a> </div> <h1 style="margin-top: 0rem;">✨ Read our DeepSeek-OCR Guide <a href="https://docs.unsloth.ai/new/deepseek-ocr">here</a>!</h1> </div>

• Thank you to Prithiv's model modifcations that enables DeepSeek-OCR fine-tuning.
• Fine-tune DeepSeek-OCR for free using our Google Colab notebook
• View the rest of our notebooks in our docs here.

---

<div align="center"> <img src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/logo.svg?raw=true" width="60%" alt="DeepSeek AI" /> </div> <hr> <div align="center"> <a href="https://www.deepseek.com/" target="_blank"> <img alt="Homepage" src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/badge.svg?raw=true" /> </a> <a href="https://huggingface.co/deepseek-ai/DeepSeek-OCR" target="_blank"> <img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-DeepSeek%20AI-ffc107?color=ffc107&logoColor=white" /> </a> </div> <div align="center"> <a href="https://discord.gg/Tc7c45Zzu5" target="_blank"> <img alt="Discord" src="https://img.shields.io/badge/Discord-DeepSeek%20AI-7289da?logo=discord&logoColor=white&color=7289da" /> </a> <a href="https://twitter.com/deepseek_ai" target="_blank"> <img alt="Twitter Follow" src="https://img.shields.io/badge/Twitter-deepseek_ai-white?logo=x&logoColor=white" /> </a> </div> <p align="center"> <a href="https://github.com/deepseek-ai/DeepSeek-OCR"><b>🌟 Github</b></a> | <a href="https://huggingface.co/deepseek-ai/DeepSeek-OCR"><b>📥 Model Download</b></a> | <a href="https://github.com/deepseek-ai/DeepSeek-OCR/blob/main/DeepSeek_OCR_paper.pdf"><b>📄 Paper Link</b></a> | <a href="https://arxiv.org/abs/2510.18234"><b>📄 Arxiv Paper Link</b></a> | </p> <h2> <p align="center"> <a href="">DeepSeek-OCR: Contexts Optical Compression</a> </p> </h2> <p align="center"> <img src="assets/fig1.png" style="width: 1000px" align=center> </p> <p align="center"> <a href="">Explore the boundaries of visual-text compression.</a> </p>

Usage

Inference using Huggingface transformers on NVIDIA GPUs. Requirements tested on python 3.12.9 + CUDA11.8：

torch==2.6.0
transformers==4.46.3
tokenizers==0.20.3
einops
addict 
easydict
pip install flash-attn==2.7.3 --no-build-isolation

from transformers import AutoModel, AutoTokenizer
import torch
import os
os.environ["CUDA_VISIBLE_DEVICES"] = '0'
model_name = 'deepseek-ai/DeepSeek-OCR'

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModel.from_pretrained(model_name, _attn_implementation='flash_attention_2', trust_remote_code=True, use_safetensors=True)
model = model.eval().cuda().to(torch.bfloat16)

# prompt = "<image>\nFree OCR. "
prompt = "<image>\n<|grounding|>Convert the document to markdown. "
image_file = 'your_image.jpg'
output_path = 'your/output/dir'

# infer(self, tokenizer, prompt='', image_file='', output_path = ' ', base_size = 1024, image_size = 640, crop_mode = True, test_compress = False, save_results = False):

# Tiny: base_size = 512, image_size = 512, crop_mode = False
# Small: base_size = 640, image_size = 640, crop_mode = False
# Base: base_size = 1024, image_size = 1024, crop_mode = False
# Large: base_size = 1280, image_size = 1280, crop_mode = False

# Gundam: base_size = 1024, image_size = 640, crop_mode = True

res = model.infer(tokenizer, prompt=prompt, image_file=image_file, output_path = output_path, base_size = 1024, image_size = 640, crop_mode=True, save_results = True, test_compress = True)

vLLM

Refer to 🌟GitHub for guidance on model inference acceleration and PDF processing, etc.<!-- -->

[2025/10/23] 🚀🚀🚀 DeepSeek-OCR is now officially supported in upstream vLLM.

uv venv
source .venv/bin/activate
# Until v0.11.1 release, you need to install vLLM from nightly build
uv pip install -U vllm --pre --extra-index-url https://wheels.vllm.ai/nightly

from vllm import LLM, SamplingParams
from vllm.model_executor.models.deepseek_ocr import NGramPerReqLogitsProcessor
from PIL import Image

# Create model instance
llm = LLM(
    model="deepseek-ai/DeepSeek-OCR",
    enable_prefix_caching=False,
    mm_processor_cache_gb=0,
    logits_processors=[NGramPerReqLogitsProcessor]
)

# Prepare batched input with your image file
image_1 = Image.open("path/to/your/image_1.png").convert("RGB")
image_2 = Image.open("path/to/your/image_2.png").convert("RGB")
prompt = "<image>\nFree OCR."

model_input = [
    {
        "prompt": prompt,
        "multi_modal_data": {"image": image_1}
    },
    {
        "prompt": prompt,
        "multi_modal_data": {"image": image_2}
    }
]

sampling_param = SamplingParams(
            temperature=0.0,
            max_tokens=8192,
            # ngram logit processor args
            extra_args=dict(
                ngram_size=30,
                window_size=90,
                whitelist_token_ids={128821, 128822},  # whitelist: <td>, </td>
            ),
            skip_special_tokens=False,
        )
# Generate output
model_outputs = llm.generate(model_input, sampling_param)

# Print output
for output in model_outputs:
    print(output.outputs[0].text)

Visualizations

<table> <tr> <td><img src="assets/show1.jpg" style="width: 500px"></td> <td><img src="assets/show2.jpg" style="width: 500px"></td> </tr> <tr> <td><img src="assets/show3.jpg" style="width: 500px"></td> <td><img src="assets/show4.jpg" style="width: 500px"></td> </tr> </table>

Acknowledgement

We would like to thank Vary, GOT-OCR2.0, MinerU, PaddleOCR, OneChart, Slow Perception for their valuable models and ideas.

We also appreciate the benchmarks: Fox, OminiDocBench.

Citation

@article{wei2025deepseek,
  title={DeepSeek-OCR: Contexts Optical Compression},
  author={Wei, Haoran and Sun, Yaofeng and Li, Yukun},
  journal={arXiv preprint arXiv:2510.18234},
  year={2025}
}