2.8.0_cuda11.8

PaddleDetection

Uses PaddleDetection (documentation).

Uses PyTorch 2.5.1, CUDA 11.8 and PaddleDetection 2.8.0.

Quick start

Inhouse registry

• Log into registry using public credentials:

  docker login -u public -p public public.aml-repo.cms.waikato.ac.nz:443 

• Pull and run image (adjust volume mappings -v):

  docker run --gpus=all --shm-size 8G --net=host \
    -v /local/dir:/container/dir \
    -it public.aml-repo.cms.waikato.ac.nz:443/paddle/paddledetection:2.8.0_cuda11.8

Docker hub

• Pull and run image (adjust volume mappings -v):

  docker run --gpus=all --shm-size 8G --net=host \
    -v /local/dir:/container/dir \
    -it waikatodatamining/paddledetection:2.8.0_cuda11.8

Build local image

• Build the image from Docker file (from within /path_to/paddledetection/2.8.0_cuda11.8)

  docker build -t paddledetection .

• Run the container

  docker run --gpus=all --shm-size 8G --net=host -v /local/dir:/container/dir -it paddledetection

  /local/dir:/container/dir maps a local disk directory into a directory inside the container

Publish images

Build

docker build -t paddledetection:2.8.0_cuda11.8 .

Inhouse registry

• Tag

  docker tag \
    paddledetection:2.8.0_cuda11.8 \
    public-push.aml-repo.cms.waikato.ac.nz:443/paddle/paddledetection:2.8.0_cuda11.8

• Push

  docker push public-push.aml-repo.cms.waikato.ac.nz:443/paddle/paddledetection:2.8.0_cuda11.8

  If error "no basic auth credentials" occurs, then run (enter username/password when prompted):

  docker login public-push.aml-repo.cms.waikato.ac.nz:443

Docker hub

• Tag

  docker tag \
    paddledetection:2.8.0_cuda11.8 \
    waikatodatamining/paddledetection:2.8.0_cuda11.8

• Push

  docker push waikatodatamining/paddledetection:2.8.0_cuda11.8

  If error "no basic auth credentials" occurs, then run (enter username/password when prompted):

  docker login

Requirements

docker run --rm \
  -it public.aml-repo.cms.waikato.ac.nz:443/paddle/paddledetection:2.8.0_cuda11.8 \
  pip freeze > requirements.txt

Permissions

When running the docker container as regular use, you will want to set the correct user and group on the files generated by the container (aka the user:group launching the container):

docker run -u $(id -u):$(id -g) -e USER=$USER ...

Caching

PaddleDetection will download pretrained models and cache them locally. To avoid having to download them constantly, you can the cache directory to the host machine:

• when running the container as current user

  -v `pwd`/cache:/.paddledetection \
  -v `pwd`/cache/visualdl:/.visualdl \

Scripts

The following additional scripts are available:

• paddledet_train - for training models (calls the /opt/PaddleDetection/tools/train.py script)
• paddledet_eval - for evaluating trained models (calls the /opt/PaddleDetection/tools/eval.py script)
• paddledet_export_model - for exporting trained models (calls the /opt/PaddleDetection/tools/export_model.py script)
• paddledet_infer - for simple inference using a trained model (calls the /opt/PaddleDetection/tools/infer.py script)