search

brandontrabucco / mineral

A minimalist reinforcement learning package for TensorFlow 2.0
https://www.btrabucco.com/
MIT License
3 stars 0 forks source link

Debugging TRPO #5

Open brandontrabucco opened 5 years ago

brandontrabucco commented 5 years ago

Training is very slow, need to check the original paper hyper parameters.

Neo-X commented 5 years ago

I find this reference helpful. https://github.com/joschu/modular_rl/blob/master/modular_rl/trpo.py