I fixed a few small bugs (shape mismatch in ppo, some import error, a bug in a function call). I also added an option to specify wandb_run_name in the command line so we can give multiple runs with the same experiment settings unique names in wandb. Additionally there are some new experiments specified for testing.
I fixed a few small bugs (shape mismatch in ppo, some import error, a bug in a function call). I also added an option to specify wandb_run_name in the command line so we can give multiple runs with the same experiment settings unique names in wandb. Additionally there are some new experiments specified for testing.