Artur Niederfahrenhorst
|
960032a15f
[RLlib][RLModules] RNNs and RLModules (#32723)
|
1 year ago |
Sven Mika
|
656fe0703c
Revert revert [RLlib] DreamerV3 Main Algo. (#36571)
|
1 year ago |
Kai Fricke
|
42e06e3948
Revert "[RLlib] DreamerV3: Main algo code and required changes to some RLlib APIs (RolloutWorker). (#35386)" (#36564)
|
1 year ago |
Sven Mika
|
8290bd112c
[RLlib] DreamerV3: Main algo code and required changes to some RLlib APIs (RolloutWorker). (#35386)
|
1 year ago |
Sven Mika
|
31ab8a7537
[RLlib] RLModule API change: If "actions" key returned from forward_inference|exploration, use actions as-is. (#36067)
|
1 year ago |
Artur Niederfahrenhorst
|
51190a13be
[RLlib] Fix missing action distribution error in policy files (#35625)
|
1 year ago |
Artur Niederfahrenhorst
|
f796635c3f
[RLlib] Add torch compile capabilities to TorchRLModule (#34640)
|
1 year ago |
Sven Mika
|
384ad04987
[RLlib] APPO+new-stack (Atari benchmark) - Preparatory PR 04 - LearnerAPI changes/tf-tracing fixes. (#34959)
|
1 year ago |
Artur Niederfahrenhorst
|
9a5ef51f5e
[RLlib] No longer return action distribution objects from RLModule's `forward_...()` methods. (#35085)
|
1 year ago |
Sven Mika
|
78b58a959a
[RLlib] Learner API: Policies using RLModules (for sampler only) do not need loss/stats/mixins. (#34445)
|
1 year ago |
kourosh hakhamaneshi
|
67706f9af9
[RLlib] Fixed bug in restoring a gpu trained algorithm (#35024)
|
1 year ago |
Sven Mika
|
adfdbbdfa2
[RLlib] APPO+new-stack (Atari benchmark) - Preparatory PR 03 - PyTorch. (#34779)
|
1 year ago |
Avnish Narayan
|
aac0fa07f4
Revert "[RLlib] Introduce experimental larger than GPU train batch size feature for torch (#34189)" (#34766)
|
1 year ago |
Artur Niederfahrenhorst
|
72268e801d
[RLlib] Introduce experimental larger than GPU train batch size feature for torch (#34189)
|
1 year ago |
kourosh hakhamaneshi
|
33ca3399c0
[RLlib][RLModule] Fix the get_state() of policies to only skip the optimizer states if learner_api is enabled (#33825)
|
1 year ago |
Artur Niederfahrenhorst
|
4b91dd0136
[RLlib] Fix saving and loading algos and policies docs example for RL Modules (#33848)
|
1 year ago |
Artur Niederfahrenhorst
|
8a80839843
[RLlib] Fix test backward compatibility test for RL Modules (#33857)
|
1 year ago |
Artur Niederfahrenhorst
|
bf0257104e
[RLlin] Add dist_inputs to action sampler fn returns in TorchPolicyV2 (#33795)
|
1 year ago |
Artur Niederfahrenhorst
|
5f40e22486
[RLlib] Add support for MultiCategorical and MultiActionDistribution for RL Modules (#32938)
|
1 year ago |
kourosh hakhamaneshi
|
836c684133
[RLlib] RLModule API does not support ONNX model export yet (#33297)
|
1 year ago |
Avnish Narayan
|
b4be536a02
[RLlib] Initialize the default view requirements when using the learner group api (#32825)
|
1 year ago |
kourosh hakhamaneshi
|
20bd70103c
[RLlib][RLModule] Implemented the logp logic in the policy (#33252)
|
1 year ago |
Sven Mika
|
cfa4dd8626
[RLlib] Issue 28428: QMix not working w/ GPUs. (#31299)
|
1 year ago |
Sven Mika
|
8e680c483c
[RLlib] gymnasium support (new `Env.reset()/step()/seed()/render()` APIs). (#28369)
|
1 year ago |
kourosh hakhamaneshi
|
f9ec2d1ae2
[RLlib] Make RLModule initialization easy (#31069)
|
1 year ago |
kourosh hakhamaneshi
|
185c8a5ba5
[RLlib] RLmodule integration POC with existing torch policies (#29868)
|
1 year ago |
Sven Mika
|
756321145d
[RLlib] Add metrics to IMPALA/APPO/PPO (prototype) to measure off-policy'ness for performed updates. (#29983)
|
1 year ago |
Sven Mika
|
23b3a599b9
[RLlib] Algorithm/Policy checkpoint overhaul and Policy Model export (in native formats). (#28166)
|
2 years ago |
Sven Mika
|
6ca0b2f8e5
[RLlib] Some minor cleanups. (#28464)
|
2 years ago |
Kai Fricke
|
7335a4908c
[tune/rllib] Ignore directory exists errors to tackle race conditions (#28401)
|
2 years ago |