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
- Runmain.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
.