- Softcode Optuna storage
- Double check that Optuna is parallelizing with multiprocessing, otherwise implement multiprocessing support
- Verify that environment reproducibility is working (seed setting)
- Plot results for different algorithms
- Add support for additional environments:
FrozenLake-v0,Taxi-v3,CliffWalking-v0 - Add target table support: eg.
Double Q-Learning - Add support for different exploration strategies
- Tune hyperparams and commit results
- Add wandb support (https://github.com/optuna/optuna-examples/blob/main/wandb/wandb_integration.py)
- Test wandb sweeps
- Test Ray Tune