facebookresearch mbrl-lib issues

facebookresearch / mbrl-lib

Library for Model Based RL

MIT License

959 stars 158 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

[Bug] MBPO loss explode on humanoid and walker (sometimes)

#137 MansonX closed 2 years ago
4
[Bug] PETS not working

#136 sofan110 closed 2 years ago
18
Feature pybullet

#135 dtch1997 closed 2 years ago
44
[Feature Request] Add support for low dimensional observations to PlaNet

#134 luisenp opened 3 years ago
0
[Feature Request] Clean up the PlaNet visualizer

#133 luisenp opened 3 years ago
0
[Feature Request] Add batch support to ModelEnv.evaluate_action_sequences

#132 luisenp opened 3 years ago
0
[Feature Request] Add batch support to CEM

#131 luisenp opened 3 years ago
0
[Feature Request] Add PyBullet implementations of envs

#130 dtch1997 closed 2 years ago
2
Changed signature of reset and sample methods to use a model state dictionary

#129 luisenp closed 3 years ago
0
Replaced usage of 'master' branch for 'main'.

#128 luisenp closed 3 years ago
0
Icem rebase

#127 luisenp closed 3 years ago
0
Removed __all__ from __init__.py files.

#126 luisenp closed 3 years ago
0
Add configurable activation function to MLP models #41

#125 JARACH-209 closed 3 years ago
3
Add configurable activation function to MLP models #41

#124 JARACH-209 closed 3 years ago
5
Support for robotics tasks in control_env + MPPI Bug fix

#123 freiberg-roman closed 2 years ago
6
Implemented Cartpole PETS version

#122 hxri closed 3 years ago
4
Fix initialization in PLANET to match tensorflow inits

#121 eugenevinitsky closed 3 years ago
0
iCEM

#120 sradicwebster closed 3 years ago
4
MPPI

#119 freiberg-roman closed 3 years ago
1
Reward setup of the HalfCheetah environment

#118 jihwan-jeong closed 3 years ago
2
[Feature Request] cartpole_continuous.py is not the standard continuous cartpole

#117 robertocalandra closed 3 years ago
6
Fixed wrong absolute value parenthesis in Hopper's termination function.

#116 luisenp closed 3 years ago
0
[Feature Request] More general reward

#115 mkolodziejczyk-piap opened 3 years ago
1
Could you please provide the data corresponding to the figures of of experiments in the paper——PETS? Because we would like to like cite PETS as one of the state of the art work in our paper, but time is limited. Thank you![Feature Request]

#114 ChengpengLi1003 closed 3 years ago
1
[Bug] Hopper-v2 crashing using MBPO

#113 sufodsia closed 3 years ago
1
Added support for per-layer weight-decay in GaussianMLP.

#112 luisenp closed 3 years ago
3
Add option to do normalization using double precision for 1-D models

#111 luisenp closed 3 years ago
0
[Bug] Normalization of model inputs causes PETS to crash on Pybullet Half Cheetah

#110 MarkSelden closed 3 years ago
3
Model functions now include metadata

#109 luisenp closed 3 years ago
1
[Feature Request] Logging of custom training metrics

#108 jan1854 closed 3 years ago
7
Fixed some type hints of ReplayBuffer

#107 jan1854 closed 3 years ago
4
Fixed a bug where a default argument to get_sequence_buffer_iterator causes a crash and added an assertion for clarity

#106 jan1854 closed 3 years ago
1
Fixed a bug where SequenceTransitionIterator.__len__() ignores max_batches_per_loop

#105 jan1854 closed 3 years ago
1
Bugfix in ModelTrainer.maybe_get_best_weights() and made the threshold for improvement configurable from ModelTrainer.train()

#104 jan1854 closed 3 years ago
1
[bug-fix] Fixed bug in code to keep best weights during model trainin…

#103 luisenp closed 3 years ago
0
[Bug] ModelTrainer.maybe_get_best_weights() does not deal properly with negative evaluation scores

#102 jan1854 closed 3 years ago
2
[Question] Model Interface: forward arguments

#101 axelbr closed 3 years ago
3
Fixed the condition of the warning in ReplayBuffer.close_trajectory()

#100 jan1854 closed 3 years ago
0
Difference in PETS implementation from the original TF version.

#99 ethanluoyc closed 3 years ago
8
[Bug] Issue with loading best weights in ModelTrainer

#98 ethanluoyc closed 3 years ago
6
[Bug] MBPO on HalfCheetah-v2 not learning

#97 sufodsia closed 3 years ago
5
Fix ReplayBuffer incorrect trajectory bookkeeping for partial trajectories

#96 luisenp closed 3 years ago
0
[Bug] ReplayBuffer.sample_trajectory() returns transitions from two different trajectories

#95 jan1854 closed 3 years ago
2
Set default model length to one

#94 luisenp closed 3 years ago
0
Changed Model.__len__ method to return 1 by default

#93 jettjaniak closed 3 years ago
0
[Bug] `__len__` method of Model should always return non-negative integer

#92 jettjaniak closed 3 years ago
0
Added sequence iterator

#91 luisenp closed 3 years ago
0
Fixed some CI errors caused by mypy update.

#90 luisenp closed 3 years ago
0
Fixed failing tests in test_common_utils and test_algorithms

#89 jettjaniak closed 3 years ago
1
Pddm

#88 freiberg-roman closed 2 years ago
13

Previous Next