kbai/ocropy
kbai
自动构建
Ocropy in a Docker container
8 次收藏下载次数: 0状态：自动构建维护者：kbai仓库类型：镜像最近更新：8 年前
让 AI 帮你使用轩辕镜像？ · 展开查看说明 · 点击收起说明
如果你使用 DeepSeek、元宝 AI、Copilot、ChatGPT 等 AI 工具，协助配置轩辕镜像、编写 docker pull 命令、修改 Docker Compose 镜像地址、配置镜像加速、排查镜像拉取失败、分析报错日志等问题，请先让 AI 阅读并遵守轩辕镜像的规则文档。
只需在 AI 对话中先发送下面这句话即可：
请先完整阅读并严格遵守以下文档中的全部规则与要求：
https://xuanyuan.cloud/agents.md
在未充分阅读并理解该文档前，不要生成任何命令、配置、修改建议、故障排查方案或技术回答。后续所有输出都必须严格以该文档中的规范为最高优先级执行。查看 agents.md 用法指南与完整示范。国内用户首推 元宝 AI、DeepSeek 的深度思考模式，不推荐豆包 AI；Cursor 等编辑器可在对话 @ 该链接，或加入 User Rules。 若 AI 无法访问外链，可 打开说明文档 复制全文粘贴。文档会随站点更新，复制内容可能过期，建议定期检查。
docker-ocropy
https://github.com/tmbdev/ocropy in a Docker container.
https://img.shields.io/docker/automated/kbai/ocropy.svg?maxAge=2592000](https://hub.docker.com/r/kbai/ocropy/) https://img.shields.io/docker/pulls/kbai/ocropy.svg?maxAge=2592000]()
This is a simple way of getting an ocropy OCR system installed.
Run it
For convenience, this repo comes bundled with a shell script run-ocropy:
> ./run-ocropy --help ./run-ocropy [--help] [--sudo] [--image <image-name>] [--working-dir <working-dir>] <cmd> [args...] --sudo Run as sudo --image <image-name> Use this image (Default: kbai/ocropy) --working-dir <working-dir> Mount this dir (Default: $PWD) <cmd> Execute this ocropy-<cmd> command [args...] Depend on the command. See ocropy docs. Commands: dewarp econf errs gpageseg gtedit hocr linegen lpred ltrain nlbin rpred rtrain results
Run with docker command
You can run ocropy from the command line:
docker run -v `pwd`:/work kbai/ocropy /ocropy/ocr***-rpred *.png
This will:
- Download the docker-ocropy container from https://hub.docker.com/r/kbai/ocropy
- Mount
[object Object]to the/workdirectory within the docker - Execute
/ocropus/ocropy/ocropus-rpredfrom within the container - On all PNG images in the current dir (
pwd) on this computer and hence/workin the container
See the Docker documentation to see how you need to use the -v flag and other flags.
You can also do training and all the other operations you'd normally run with an ocropy installation.
Models
By default, docker-ocropy ships with the models
ocropy/en-default and
ocropy/fraktur. These are found in
/ocropy/models, i.e. use
shdocker run --rm -it kbai/ocropy ocropus-rpred -m en-default.pyrnn.gz ... # or docker run --rm -it kbai/ocropy ocropus-rpred -m /ocropy/models/en-default.pyrnn.gz ...
Train it
Use the ocrotrain.sh script to see how to train the engine:
docker run -v "$PWD:/work" kbai/ocropy ./ocrotrain.sh
This will download the https://storage.googleapis.com/tmbdev-public/uw5-100.tgz dataset
and run ocropus-rtrain on it.
Build it
If you want to adapt the Dockerfile and rebuild the image, you can do so with
docker build [-t '<username>/<reponame>'] .
While optional for local use, choosing a username/reponame combo makes it easier to manage
镜像拉取方式
您可以使用以下命令拉取该镜像。请将 <标签> 替换为具体的标签版本。如需查看所有可用标签版本，请访问 标签列表页面。
DockerHub 原生拉取命令
docker pull kbai/ocropy:<标签>
镜像拉取常见问题
功能
错误码
用户好评
来自真实用户的反馈，见证轩辕镜像的优质服务
oldzhang
运维工程师
Linux服务器
5
"Docker访问体验非常流畅，大镜像也能快速完成下载。"