updated readme

Mujoco test cases

adding Mujoco support, refactoring: unifying video recording API, adding tests

more modularity and added GRPO support

all tests passed

added test_mlp.py

grpo seems to train, but not optimal

PPO trains

trains

nightly

get_agent now takes env

print -> debug print

first time tests passed

changed done and reward to [num_envs] instead of [num_envs, 1]

added training tests