Software Engineering for Research, originally from cvanea's: https://github.com/cvanea/ser/ for the CDT for Health Data Science software engineering module.
conda create -n ser python=3.9
conda activate ser
pip install -e .
There are instructions on the PyTorch website how to install the PyTorch versions that are correct for your system...
https://pytorch.org/get-started/locally/
Ideally install using pip.
conda activate ser
ser --help
Run run_cli.py with debug config "run_cli_train" or try "Python Debugger: Current File with Arguments" to define your own entry to cli.py.
You can also try running jupyter_eg.ipynb to try debugging in jupyter.
- Note usually would add ".vscode" to .gitignore so not uploaded to repoitory as your personal config.