This PR includes code to tune PC and SQIL using optuna on the SLURM cluster.
It does not use the tune.py or parallel.py scripts as they were not compatible with running on SLURM.
The included jupyter notebook can be used to aggregate the benchmark runs in a table for further analysis.
It has to be executed in the folder containing the output of the tuning process.
The tuning results are on the NAS: /nas/ucb/maximilian/imitation/tuning/
This PR includes code to tune PC and SQIL using optuna on the SLURM cluster.
It does not use the tune.py or parallel.py scripts as they were not compatible with running on SLURM.
The included jupyter notebook can be used to aggregate the benchmark runs in a table for further analysis. It has to be executed in the folder containing the output of the tuning process. The tuning results are on the NAS:
/nas/ucb/maximilian/imitation/tuning/