Speeding up training - Montezuma's revenge - gym-ai


#1

Hi All,

I am trying to train an agent to play Montezuma’s revenge making use of hierarchical reinforcement learning. I am very new to the area and currently training it on the GPU of my laptop. Due to the amount of computation it needs for each step, it’s extremely slow. Any suggestions on speeding it up so I can see the results quicker, like a machine learning platform on cloud that works well with gym?

Thanks a lot for your help in advance!