This code implements the training loop for the Mujoco set of experiments.

To run the code launch on the terminal the command (code runnable on UNIX systems only):

python3 trainer_fetch.py --algo <representation_id> --env_id <environment_id> --z_dim <z_dim> --seed <seed>

<representation_id>  0: "LinearComb", 1: "ConCat", 2: "Curl", 3: "MMM", 4: "GMC", 5: "AMDF", 6: "CORAL"
<environment_id>     0: "Ant-v5", 1: "HalfCheetah-v5", 2: "Hopper-v5", 3: "Humanoid-v5", 4: "Walker2d-v5", 5: "InvertedPendulum-v5"
<z_dim>              dimension of the latent space representation, default is 64.
<seed>               sets the seed, default is 0.


During training, the best performing model so far is saved in the folder 'saved_assets'



