Files
2026-07-13 10:09:03 +00:00

8.0 KiB
Raw Permalink Blame History

Note

本文档由 WeHub 基于上游 README 翻译整理,属于社区翻译,非官方中文文档。
English · 原始项目 · 上游 README
原作者、版权与许可证归属以原始项目及本仓库 LICENSE 文件为准。

DeepSeek AI

📥 模型下载 | 📄 论文链接 | 📄 Arxiv 论文链接 |

DeepSeek-OCR:上下文光学压缩(Contexts Optical Compression

探索视觉-文本压缩的边界。

发布

  • [2026/01/27]🚀🚀🚀🚀🚀🚀 我们发布 DeepSeek-OCR2
  • [2025/10/23]🚀🚀🚀 DeepSeek-OCR 现已在上游 vLLM. 中获得官方支持。感谢 vLLM 团队的帮助。
  • [2025/10/20]🚀🚀🚀 我们发布 DeepSeek-OCR,这是一个从以 LLM 为中心的视角研究视觉编码器作用的模型。

目录

安装

我们的环境为 cuda11.8+torch2.6.0。

  1. 克隆本仓库并进入 DeepSeek-OCR 文件夹
git clone https://github.com/deepseek-ai/DeepSeek-OCR.git
  1. Conda
conda create -n deepseek-ocr python=3.12.9 -y
conda activate deepseek-ocr
  1. 依赖包
  • 下载 vllm-0.8.5 whl
pip install torch==2.6.0 torchvision==0.21.0 torchaudio==2.6.0 --index-url https://download.pytorch.org/whl/cu118
pip install vllm-0.8.5+cu118-cp38-abi3-manylinux1_x86_64.whl
pip install -r requirements.txt
pip install flash-attn==2.7.3 --no-build-isolation

注意: 若你希望 vLLM 与 Transformers 代码在同一环境中运行,则无需担心如下安装错误:vllm 0.8.5+cu118 requires transformers>=4.51.1

vLLM 推理

  • VLLM:

注意: 请在 DeepSeek-OCR-master/DeepSeek-OCR-vllm/config.py 中修改 INPUT_PATH/OUTPUT_PATH 及其他设置

cd DeepSeek-OCR-master/DeepSeek-OCR-vllm
  1. 图像:流式输出
python run_dpsk_ocr_image.py
  1. PDF:并发约 2500 tokens/sA100-40G
python run_dpsk_ocr_pdf.py
  1. 基准测试批量评估
python run_dpsk_ocr_eval_batch.py

[2025/10/23] 上游 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)

Transformers 推理

  • Transformers
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'

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)

或者你可以

cd DeepSeek-OCR-master/DeepSeek-OCR-hf
python run_dpsk_ocr.py

支持模式

当前开源模型支持以下模式:

  • 原生分辨率:
    • Tiny: 512×512 64 vision tokens
    • Small: 640×640 100 vision tokens
    • Base: 1024×1024 256 vision tokens
    • Large: 1280×1280 400 vision tokens
  • 动态分辨率
    • Gundam: n×640×640 + 1×1024×1024

提示词示例

# document: <image>\n<|grounding|>Convert the document to markdown.
# other image: <image>\n<|grounding|>OCR this image.
# without layouts: <image>\nFree OCR.
# figures in document: <image>\nParse the figure.
# general: <image>\nDescribe this image in detail.
# rec: <image>\nLocate <|ref|>xxxx<|/ref|> in the image.
# '先天下之忧而忧'

可视化

致谢

我们感谢 Vary, GOT-OCR2.0, MinerU, PaddleOCR, OneChart, Slow Perception 提供的宝贵模型与思路。

我们也感谢以下基准测试:Fox, OminiDocBench.

引用

@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}
}