weidler / RLaSpa

Reinforcement Learning in Latent Space
MIT License
5 stars 1 forks source link

Batch Learning #26

Closed weidler closed 5 years ago

weidler commented 5 years ago

Works reasonably for all non-pixel tasks. For pixel tasks its hella slow, I don't know if thats my fault. @HansBambel should look into that since he has the experience with this Pixel stuff.

I had to change/remove/move some things from the learners to the agent. E.g. representing the action as a one-hot-vector.