• (For docker) rootless/rootful docker

1. Change variables in ./docker/docker-compose.yml as needed. Especially, you may substitute pytorch-template in the file with your project name.

2. Add scripts into ./docker/rootless/Dockerfile as needed. It will works without any changes.

3. Change variables in pyproject.toml as needed. Especially, you may change name and authors in the file.

4. To build your Docker image, run the following command in the directory containing the docker-compose.yml file:

     docker compose build
   

5. Run a docker container via the following command

     docker compose up -d
   

6. Run bash inside of the container

     docker compose exec <project_name> /bin/bash
   

7. Change variables in pyproject.toml as needed

   • e.g., name, version, description, and authors

8. Remove unrequired dependencies in pyproject.toml

9. Create a vertual environment and install dependencies by

     poetry install
   

10. Add required dependencies by

      poetry add <library name>
    

11. Run the following command to verify that your installation was successful.

      poetry run python scripts/installation_verification.py
    

Note that all changes in the docker container will be deleted when you exit docker container.

To enable GPU calculations, you can override docker-compose.yml with docker-compose.gpu.yml by creating ./docker/.env file as follows:

  echo COMPOSE_FILE=docker-compose.yml:docker-compose.gpu.yml >> ./docker/.env

If you're unable to use rootless Docker, you'll need to override the default settings with docker-compose.rootful.yml. Also, you may add scripts into ./docker/rootful/Dockerfile.

1. Run a docker container

     docker compose up -d
   

2. Run bash inside of the container

     docker compose exec <project_name> /bin/bash
   

3. Run docker/run_jupyter.sh

     ./docker/run_jupyter.sh
   

4. When you write and run your scripts, you need to choose the kernel project_name

1. For prototyping, make notebooks/local directory and add jupyter notebook into the directory.
2. For developing, add source files into src directory
3. For testing, add test codes into test directory
4. For evaluation, add script files into scripts directory