HumanCompatibleAI/imitation
Clean PyTorch implementations of imitation and reward learning algorithms
PythonMIT
Issues
- 1
"The observation_space and action_space were not given, can't verify new environments" when loading expert policy
#854 opened by boshi-an - 1
Unable to load FeedForward32Policy -- multiple values for keyword argument 'net_arch'
#857 opened by glolichen - 0
- 3
The observation_space and action_space were not given, can't verify new environments
#819 opened by kavinwkp - 0
GAIL tensorboard logging does not appear to work
#853 opened by eufrizz - 1
- 1
Problem in Passing Along Custom Gym Env Constructor Parameters in make_vec_env
#850 opened by christianjcc - 1
Serialize Dataset Save Not Working
#851 opened by alexpalms - 0
Preference based Reinforcement Learning applies a "recurrent reward network" for solving a POMDP problem
#848 opened by CAI23sbP - 0
- 5
Got an unexpected keyword argument 'use_sde' when passing behavioural cloning policy to PPO from SB3
#781 opened by JkAcktuator - 1
RewardNetwork predict_processed doesn't work without next_state and done
#836 opened by gustavodemari - 0
- 0
difference between Step and Sample in Dagger
#846 opened by lwizard1999 - 1
Device conflict when training BC
#843 opened by IsaacSheidlower - 2
Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!
#820 opened by kavinwkp - 1
segmentation fault while loading policy
#845 opened by chenyangkang - 1
abort and kernel crash with OMP libomp.dylib
#844 opened by chenyangkang - 4
Utilizing Expert Data (.npz) in format of SB3.
#777 opened by azafar1991 - 2
- 0
Rewards visualization
#840 opened by Rajesh-Siraskar - 0
[Question] Training in Custom Environment
#839 opened by KingsDevs - 0
[Question] Reward net transfer
#838 opened by risufaj - 2
prob_true_act > 1 with PPO and BC?
#762 opened by ThomasRochefortB - 0
Remove FloatReward when next SB3 is released
#794 opened by ZiyueWang25 - 0
Hierarchical Behavior Cloning
#834 opened by Dhanushvarma - 3
Example notebook "1_train_bc.ipynb" gives "Namespace not found error" for "seals" module
#816 opened by Rajesh-Siraskar - 8
Ensure all tutorials work as expected
#763 opened by ernestum - 1
ReadTimeout during hyperparameter tuning
#804 opened by ZiyueWang25 - 3
Observation space not supported when getting expert trajectories in tutorial 7
#788 opened by camberg23 - 1
- 1
error with setting up reward net
#809 opened by spearsheep - 2
DAgger Demo Code not working
#814 opened by nil123532 - 1
Tensorboard Logging
#818 opened by mertalbaba - 1
IQ-Learn
#813 opened by azafar1991 - 1
- 1
Run time Error when run quickstart.py
#825 opened by Charles-Lim93 - 4
Examples Don't work
#815 opened by AdvanceXplorer - 0
Improve pipeline speed and abort early
#789 opened by ernestum - 1
tests/algorithms/test_sqil.py::test_sqil_performance_continuous[DDPG] failure
#791 opened by ZiyueWang25 - 0
Fix flaky SQIL test
#807 opened by AdamGleave - 1
- 0
Shorten timeout and ensure the notebook can still show desired improvement.
#793 opened by ZiyueWang25 - 0
- 2
- 0
- 2
Add CLI for SQIL
#780 opened by AdamGleave - 0
Consider allowing integer reward in trajectories
#783 opened by PavelCz - 2
Generalize SQIL to work with other off-policy algos
#767 opened by jas-ho - 1