Set up and maintain a machine learning project in Tensorflow with a single script.

Inspired by create-react-app and create-ml-app. Made by radi-cho.

Make sure Conda is available on your device. Alternatively, if you don't need environment management, you can just use main.py as a template or edit tf_app.sh per your needs.

git clone https://github.com/radi-cho/create-tf-app.git tf_app
cd tf_app
sh ./tf_app.sh {create_env|run|lint|run_tensorboard}

• create_env - Create a Conda environment (named tf_app by default) and install necessary packages.
• run - Run main.py in the appropriate environment and pass any arguments to it.
• lint - Run flake8 linting for all .py files unless specific filenames are be provided.
• run_tensorboard - Run Tensorboard to visualize records in the ./logs directory.

You can use an environment variable - TF_APP_ENV_NAME - to create multiple projects or set a specific name.

TF_APP_ENV_NAME=tf_app2 sh ./tf_app.sh {create_env|run|lint|run_tensorboard}

If you need to edit your environment or execute specific commands in it, you can always enter it with conda activate tf_app.