Sven Mika
53206dd440
[RLlib] CQL BC loss fixes; PPO/PG/A2|3C action normalization fixes ( #16531 )
2021-06-30 12:32:11 +02:00
Gerges Dib
f8cf4a1985
[RLlib] Fixed import tensorflow when module not available ( #16171 )
2021-06-04 10:07:59 +02:00
Sven Mika
c9d220bcda
[RLlib] Upgrade RLlib regression test scripts to new testing tool - RLlib release logs for 1.4. ( #16080 )
2021-06-01 17:39:18 +02:00
Sven Mika
839fc59224
[RLlib] CQL TensorFlow support ( #15841 )
2021-05-18 11:10:46 +02:00
Sven Mika
bc09e75b78
[RLlib] Fix 3 flakey test cases. ( #15785 )
2021-05-16 12:20:33 +02:00
Michael Luo
4cbe13cdfd
[RLlib] CQL loss fn fixes, MuJoCo + Pendulum benchmarks, offline-RL example script w/ json file. ( #15603 )
...
Co-authored-by: Sven Mika <sven@anyscale.io>
Co-authored-by: sven1977 <svenmika1977@gmail.com>
2021-05-04 19:06:19 +02:00
Sven Mika
e973b726c2
[RLlib] Support native tf.keras.Models (part 2) - Default keras models for Vision/RNN/Attention. ( #15273 )
2021-04-30 19:26:30 +02:00
Sven Mika
bb8a286cbc
[RLlib] Support native tf.keras.Model (milestone toward obsoleting ModelV2 class). ( #14684 )
2021-04-27 10:44:54 +02:00
Sven Mika
cecfc3b43b
[RLlib] Multi-GPU support for Torch algorithms. ( #14709 )
2021-04-16 09:16:24 +02:00
Sven Mika
b267f1f1ba
[RLlib] Add support for Int-Box action spaces. ( #15012 )
2021-04-11 13:16:01 +02:00
Jack Parsons
3df7a010b1
[RLlib] Fixing conv filters config for ComplexInputNetwork ( #14749 )
2021-03-24 16:15:36 +01:00
Sven Mika
04bc0a9828
[RLlib] Remove all non-trajectory view API code. ( #14860 )
2021-03-23 09:50:18 -07:00
Sven Mika
e7557ae433
[RLlib] Issue 13132: DQN does not update target net after restore ( #14838 )
2021-03-23 08:30:37 +01:00
Kai Fricke
be30b784a4
Amend #14308 (fix for post_fcnet_hiddens) ( #14354 )
2021-03-22 15:44:18 +01:00
Sven Mika
69202c6a7d
[RLlib] Obsolete usage tracking dict via sample batch. ( #13065 )
2021-03-17 08:18:15 +01:00
Sven Mika
ee4b6e7e3b
[RLlib] Unity3D example broken due to change in ML-Agents API. Attention-net prev-n-a/r. Attention-wrapper works with images. ( #14569 )
2021-03-12 18:27:25 +01:00
Kai Fricke
d9e5d5f47a
[RLlib] Cast fcnet_hiddens to list for DQN models (list vs tuple mismatch error) ( #14308 )
2021-02-25 08:06:08 +01:00
Sven Mika
95ef04b71a
[RLlib] Implement TorchPolicy.export_model. ( #13989 )
2021-02-22 17:09:40 +01:00
Sven Mika
9ac731558b
[RLlib] Unify fcnet initializers for the value output layer (std=1.0 in torch, but 0.01 in tf). ( #13733 )
2021-02-02 18:42:49 +01:00
Sven Mika
0a0d9183fe
[RLlib] Trajectory view API example script (enhancements and tf2 support). ( #13786 )
2021-02-02 18:42:18 +01:00
Sven Mika
52c94b7ee9
[RLlib] Allow SAC to use custom models as Q- or policy nets and deprecate "state-preprocessor" for image spaces. ( #13522 )
2021-02-02 13:05:58 +01:00
Sven Mika
4bc257f4fb
[RLlib] Fix custom multi action distr ( #13681 )
2021-01-28 19:28:48 +01:00
cathrinS
d4ef5c5993
[RLlib] Atari-RAM-Preprocessing, unsigned observation vector results in a false preprocessed observation ( #13013 )
2021-01-28 12:07:00 +01:00
Jan Blumenkamp
964689b280
[RLlib] Fix bug in ModelCatalog when using custom action distribution ( #12846 )
...
* return tuple returned from _get_multi_action_distribution when using custom action dict
* Always return dst_class and required_model_output_shape in _get_multi_action_distribution
* pass model config to _get_multi_action_distribution
2021-01-25 12:42:39 +01:00
Saeid
d11e62f9e6
[RLlib] Fix problem in preprocessing nested MultiDiscrete ( #13308 )
2021-01-21 16:36:11 +01:00
Sven Mika
56878221ed
[RLlib] Redo: Make TFModelV2 fully modular like TorchModelV2 (soft-deprecate register_variables, unify var names wrt torch). ( #13363 )
2021-01-14 14:44:33 +01:00
Sven Mika
d49c3fae0b
[RLlib] Trajectory View API: Atari framestacking. ( #13315 )
2021-01-13 08:53:34 +01:00
Kai Fricke
25f10a947a
Revert "[RLlib] Make TFModelV2 behave more like TorchModelV2: Obsolete register_variables. Unify variable dicts. ( #13339 )" ( #13361 )
...
This reverts commit e2b2abb88b
2021-01-12 12:33:57 +01:00
Sven Mika
e2b2abb88b
[RLlib] Make TFModelV2 behave more like TorchModelV2: Obsolete register_variables. Unify variable dicts. ( #13339 )
2021-01-11 22:42:30 +01:00
Sven Mika
5d50d37f45
[RLlib] Issue 13330: No TF installed causes crash in ModelCatalog.get_action_shape() ( #13332 )
2021-01-11 13:19:46 +01:00
Sven Mika
6f342a2221
[RLlib] Preparatory PR for: Documentation on Model Building. ( #13260 )
2021-01-08 10:56:09 +01:00
Sven Mika
bcaff63909
[RLlib] SquashedGaussians should throw error when entropy or kl are called. ( #13126 )
2021-01-07 15:07:35 +01:00
Sven Mika
9eba1871bb
[RLlib] Support easy use_attention=True flag for using the GTrXL model. ( #11698 )
2021-01-01 14:06:23 -05:00
Sven Mika
8726521604
[RLlib] JAXPolicy prep PR #2 (move get_activation_fn (backward-compatibly), minor fixes and preparations). ( #13091 )
2020-12-30 22:30:52 -05:00
Sven Mika
391cdfae8c
[RLlib] Trajectory view API docs. ( #12718 )
2020-12-30 17:32:21 -08:00
Sven Mika
c524f86785
[RLlib] BC/MARWIL/recurrent nets minor cleanups and bug fixes. ( #13064 )
2020-12-27 09:46:03 -05:00
Sven Mika
a5318961de
[RLlib] Preprocessor fixes (multi-discrete) and tests. ( #13083 )
2020-12-26 20:14:36 -05:00
Sven Mika
99ae7bae05
[RLlib] JAXPolicy prep. PR #1 . ( #13077 )
2020-12-26 20:14:18 -05:00
Sven Mika
d5604eaba3
[RLlib] Attention nets PyTorch support and cleanup (using traj. view API). ( #12029 )
2020-12-21 18:38:34 -08:00
Sven Mika
b2bcab711d
[RLlib] Attention Nets: tf ( #12753 )
2020-12-20 20:22:32 -05:00
Sven Mika
124c8318a8
[RLlib] Fix broken test_distributions.py (test_categorical) ( #12915 )
2020-12-17 17:44:26 -06:00
Sven Mika
abb1eefdc2
[RLlib] Issue 12483: Discrete observation space error: "ValueError: ('Observation ({}) outside given space ..." when doing Trainer.compute_action. ( #12787 )
2020-12-11 22:43:30 +01:00
Sven Mika
99c81c6795
[RLlib] Attention Net prep PR #3 . ( #12450 )
2020-12-07 13:08:17 +01:00
Sven Mika
3f4bc16276
[RLlib] Add a minimal JAX ModelV2 (FCNet) to RLlib. ( #12502 )
2020-12-03 15:51:30 +01:00
Sven Mika
19c8033df2
[RLlib] Fix most remaining RLlib algos for running with trajectory view API. ( #12366 )
...
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* WIP.
* LINT and fixes.
MB-MPO and MAML not working yet.
* wip
* update
* update
* rmeove
* remove dep
* higher
* Update requirements_rllib.txt
* Update requirements_rllib.txt
* relpos
* no mbmpo
Co-authored-by: Eric Liang <ekhliang@gmail.com>
2020-12-01 17:41:10 -08:00
Sven Mika
3ad9365e1d
[RLlib] Attention Net prep PR #2 : Smaller cleanups. ( #12449 )
2020-12-01 08:21:45 +01:00
Sven Mika
592c161032
[RLlib] Issue 12118: LSTM prev-a/r should be separately configurable. Fix missing prev-a one-hot encoding. ( #12397 )
...
* WIP.
* Fix and LINT.
2020-11-25 11:27:46 -08:00
Sven Mika
62c7ab5182
[RLlib] Trajectory view API: Enable by default for PPO, IMPALA, PG, A3C (tf and torch). ( #11747 )
2020-11-12 16:27:34 +01:00
Michael Luo
59ccbc0fc7
[RLlib] Model Annotations: Tensorflow ( #11964 )
2020-11-12 12:18:50 +01:00
Michael Luo
b2984d1c34
[RLlib] Model Annotations to Torch Models ( #9749 )
2020-11-12 12:16:12 +01:00
