adeptRL
adeptRL copied to clipboard
Reinforcement learning framework to accelerate research
Results
20
adeptRL issues
Sort by
recently updated
recently updated
newest added
would be nice to replicate some of the work in the papers where convLSTM, RMC, RMA are used. currently action space is all 1d outputs. 84x84 outputs are treated as...
* hide dependency log messages * module-separated loggers
* pretty sure SimpleEnvManager does not work for starcraft 2 * can reproduce by swapping SimpleEnvManager into replay_gen_sc2.py