Python 3.10.10
conda env create -f environment.yaml
pip install -r requirements.txt
For the basic model, adjust various parameters in the config/testfile.yaml, and use:
python read_config.py config/testfile.yaml --run-id=<run-identifier>
You can use:
python read_congig.py --help
view all available command line parameters.
You can resume the training and force start it through parameters '--resume' and '--force'.
python read_config.py config/testfile.yaml --run-id=test_1 --resume
python read_config.py config/testfile.yaml --run-id=test_1 --force
Enable tune parameters like file in the config/tune.yaml to enable ray.tune:
python read_config.py config/tune.yaml --run-id=<run-identifier>
Tensorboard can be used:
tensorboard --logdir=path/you/set/<run-identifier>