Skip to content

Latest commit

 

History

History
16 lines (9 loc) · 1.1 KB

ReadMe.md

File metadata and controls

16 lines (9 loc) · 1.1 KB

Docker container for Layout Detection & OCR exploration.

Copy SkellefteaKraft.pth , judgement_prima_table_layout_modelv3.pth models to deep-learning-tools/layout-detection-ocr-explorer/troubleshoot/models folder.

  1. docker build . -t layout-detection-ocr-explorer:1.0-latest

  2. docker run -p 8888:8888 --name layout-detection-ocr-explorer -v <root_dir_full_path>/deep-learning-tools/layout-detection-ocr-explorer/troubleshoot:/home/jovyan/work -it layout-detection-ocr-explorer:1.0-latest

  3. if already container exists with this name , remove using docker rm layout-detection-ocr-explorer

  4. docker exec -it layout-detection-ocr-explorer /bin/bash

Once the container is up , copy the URL in below pattern from container logs to Google Colab "Connect to local runtime" dialog and replace 127.0.0.1 with localhost. Should be able to connect Google colab to local docker container with all dependencies.
Open the jupyter notebook URL directly to browser and use the notebook for troubleshooting.

Example: http://127.0.0.1:8888/lab?token=8d55859d55c83fdc6813567220ca43ac471a7f915b73c299