In the repository we have implemented Random Search, Bayesian Optimiztion, Hyperband and HpBandSter (BOHB).
install Poetry (pip install poetry) for
python 3.7.10;
Run poetry install to install the correct dependencies.
Poetry will create a .venv folder and put everything there.
Download the Intel Data Lab Dataset and set the "data_path" argument at read_and_preprocess_data accordingly.
make sure you activate the virtualenv that was created by poetry beforehand via this command:
source $(poetry env info --path)/bin/activateAfter configuring the desired model and optimization algorithm in optimization.py, lean back and
run the code via this command (note that some of the algorithms may take very
long to finish (particularly so with the RNN as it isn't able to paralellize)):
python optimization.py