AutoML Framework

Environment

Python 3.10.10

Use:

conda env create -f environment.yaml
pip install -r requirements.txt

High-level Description

AutoTorch

How to use

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>