Improving Policy Exploitation in Online Reinforcement Learning with Instant Retrospect Action

=======

GPU(+CPU) version

$ conda install -c pytorch -c nvidia faiss-gpu=1.8.0

Train IRA

$ python main.py --env=HalfCheetah-v3 --device=cuda:0 --seed= --beta=1.0 --k=10 --warmup_timestamps=4000 --action_buffer_size=200000 --policy_freq=1 --start_timesteps=10000 --policy=IRA