ryanjulian / rllab

rllab is a framework for developing and evaluating reinforcement learning algorithms, fully compatible with OpenAI Gym.
Other
1 stars 21 forks source link

Drawing window with glfw throws segmentation fault #110

Closed jonashen closed 6 years ago

ryanjulian commented 6 years ago
jonashen commented 6 years ago

It has to do with the incomplete/buggy implementation of asynchronous rollout plotting in Theano's batch_polopt. I'm going to revert the code to synchronous plotting because it works and won't break the integration branch.