Skip to content

Latest commit

 

History

History
12 lines (8 loc) · 682 Bytes

README.md

File metadata and controls

12 lines (8 loc) · 682 Bytes

RetinaNetDocker-

A Nvidia docker container environment for training and executing RetinaNet with keras.

To Use Run the following commands

  • sudo nvidia-docker build -f Dockerfile -t retinanet .

  • sudo nvidia-docker run -it -d -p 0.0.0.0:6006:6006 -p 0.0.0.0:8888:8888 -v {data path}:/data/retinanet

  • nvidia-docker exec -it {docker id} bash

  • Replace the classes in the keras_retinanet pascal preprocessor file with custom classes

  • keras_retinanet/bin/train.py pascal {/data/path to data}