类库 › MinerU-Popo
opendatalab

opendatalab/MinerU-Popo

MinerU-Popo是一个轻量级通用OCR后处理框架,旨在将页面级OCR解析转化为文档级语义结构。它通过4B模型执行表格/文本截断、标题层级及图文关联分析,解决跨页几何不连续和长文档处理难题,显著提升文档树结构构建的准确性与一致性。

技术栈

根目录 python

测试

pytest

网络

Requests
查看全部依赖 (150)

依赖

Jinja2 MarkupSafe NumPy PyJWT PyMuPDF PyYAML Pydantic Pygments accelerate aiohappyeyeballs aiosignal akshare albumentations annotated-doc annotated-types anthropic anyio astor av baostock beautifulsoup4 blake3 boto3 botocore bs4 cbor2 certifi charset-normalizer click cloudpickle contourpy cuda-bindings cuda-pathfinder cupy-cuda12x curl_cffi cycler diskcache dnspython docker docstring_parser einops email-validator exceptiongroup fastapi-cli fastapi-cloud-cli fastrlock filelock fonttools frozenlist fsspec gguf gradio_pdf h11 hf-xet httpcore httptools httpx httpx-sse huggingface_hub idna ijson iniconfig interegular jmespath kiwisolver lark markdown-it-py matplotlib mcp mdurl mpmath msgspec multidict networkx ninja nvidia-cublas-cu12 nvidia-cuda-cupti-cu12 nvidia-cuda-nvrtc-cu12 nvidia-cuda-runtime-cu12 nvidia-cudnn-cu12 nvidia-cudnn-frontend nvidia-cufft-cu12 nvidia-cufile-cu12 nvidia-curand-cu12 nvidia-cusolver-cu12 nvidia-cusparse-cu12 nvidia-cusparselt-cu12 nvidia-ml-py nvidia-nccl-cu12 nvidia-nvjitlink-cu12 nvidia-nvshmem-cu12 nvidia-nvtx-cu12 openai-harmony opencv-python partial-json-parser pdftext pillow pluggy prometheus-fastapi-instrumentator propcache psutil py4j pybase64 pycountry pydantic-extra-types pydantic_core pylatexenc pyparsing pyspark python-dateutil python-json-logger qwen-vl-utils ray regex rich rich-toolkit rignore s3transfer safetensors sentencepiece sentry-sdk shellingham six soupsieve sse-starlette supervisor sympy tenacity thop tiktoken tokenizers tomli torch torchvision tqdm transformers triton typer typing-inspection typing_extensions ultralytics ultralytics-thop urllib3 uvloop watchfiles xinghe yarl zai zai-sdk zss

评论

Home - Wiki
Copyright © 2011-2026 iteam. Current version is 2.155.2. UTC+08:00, 2026-06-07 13:40
浙ICP备14020137号-1 $Map of visitor$