I am a fresh new Docker learner.
Build a docker image and deploy it to train an easy GAN model.
This exercise is not about how good the model is.
- The purpose of the exercise is to build and deploy a docker image for python training.
- User can export the training result to the local device.
- Build a Dockerfile script.
- Build docker image with
conda
environment and specify environment with.yml
file.- Activate the environment.
- Build with shell script.
- Run the
train_gan.py
script with the environment, invoking with run.sh script. - Assign local path in train file and shell script.
Interacting with sql database software. (see folder)
- Parse data and save data in the container.
- Export the container as a docker image.