search

Xingyu-Lin / mbpo_pytorch

A pytorch reprelication of the model-based reinforcement learning algorithm MBPO
150 stars 38 forks source link

test finite loop problem & epoch lenght multiplied bug fix #5

Closed songminjae closed 2 years ago

songminjae commented 2 years ago
  1. epoch length is multiplied twice -> fix, related with #3 + by testing with Hopper & Walker #4 solved
  2. for some env that could not return done for specific epoch may suffer from infinite loop -> fix
  3. minor argument fix for better interpretation