An open-source document parsing model with SOTA accuracy and speed, using an RSI loop during post-training.
[project]
name = "somark-oxr"
version = "0.1.0"
description = "Optical Everything Recognition — high-performance document parsing"
readme = "README.md"
license = "LicenseRef-OXR"
license-files = ["LICENSE", "NOTICE"]
requires-python = ">=3.10"
dependencies = [
"typer",
"fastapi",
"uvicorn[standard]",
"python-multipart",
"openai",
"sopdf>=0.2.0",
"opencv-python",
"pydantic-settings",
"PyYAML",
"httpx",
]
[project.scripts]
oxr = "oxr.cli.main:app"
[project.urls]
Homepage = "https://github.com/SoMarkAI/OXR"
Repository = "https://github.com/SoMarkAI/OXR"
Issues = "https://github.com/SoMarkAI/OXR/issues"
"Hugging Face" = "https://huggingface.co/SoMarkAI/OXR-1.0"
ModelScope = "https://modelscope.cn/models/SoMark/OXR-1.0"
[build-system]
requires = ["setuptools>=77.0.0"]
build-backend = "setuptools.build_meta"
[tool.setuptools.packages.find]
include = ["oxr*"]
[tool.setuptools.package-data]
"oxr.server" = ["demo/**/*", "demo/*"]