crivella1/ocr_translate repository overview

⁠Description

This image will run a fully working version of the ocr_translate backend server⁠.

It is a self-hosted backend server meant for running OCR and translation of images. The server is designed to work in combo with this browser extension⁠ (also on github https://github.com/Crivella/ocr_extension⁠), that will grab images from an active tab and in-place translate them by overlaying textboxes on top of it.

⁠Running

docker run --name CONTAINER_NAME -v OCT_MODELS:/models -v OCT_DB_DATA:/db_data -v OCT_PLUGINS_DATA:/plugin_data -p SERVER_PORT:4000 -d crivella1/ocr_translate:latest

⁠GPUs

For now only CUDA is supported. Add --gpus all to the previous command and set the environment variable DEVICE to cuda

⁠List of docker-specific variables

The rest of the project variables are described on the Documentation⁠

⁠Volumes

These can either be native docker volumes or bind mounted