Amog Kamsetty and GitHub
f9a99f20dd
Revert "Re-Revert "[Core] zero-copy serializer for pytorch ( #12344 )" ( #12478 )" ( #12515 )
...
This reverts commit 3f22448834 .
2020-11-30 19:05:55 -08:00
SangBin Cho and GitHub
8223a33bff
[Logging] Log rotation on all components ( #12101 )
...
* In Progress.
* Done.
* Fix the issue.
* Add wait for condition because logs are not written right away now.
* debug string.
* lint.
* Fix flaky test.
* Fix issues.
* Fix test.
* lint.
2020-11-30 19:03:55 -08:00
Max Fitton and GitHub
2708b3abbc
[Dashboard][Bug] Fix duplicate node total rows in dashboard ( #12410 )
...
* Fix duplicate node total rows in dashboard by changing the react key of the NodeTotalRow component from the node IP to the node ID (node IP can be duplicated in the case of docker).
* simplify a piece of test code and fix a flaky time out
* lint
2020-11-30 18:43:09 -08:00
Ian Rodney and GitHub
e422ace053
[serve] Create CurrentState & GoalState ( #12369 )
2020-11-30 17:34:30 -08:00
Eric Liang and GitHub
234df9091e
[autoscaler] Try to improve the request_resources() documentation ( #12465 )
2020-11-30 16:03:30 -08:00
Richard Liaw and GitHub
9ce7ad17fd
[tune] remove some bottlenecks in trialrunner ( #12476 )
2020-11-30 14:54:25 -08:00
Ian Rodney and GitHub
f5fe3794c8
[Docker] Uninstall Typing ( #12500 )
2020-11-30 14:12:57 -08:00
Siyuan (Ryans) Zhuang and GitHub
3f22448834
Re-Revert "[Core] zero-copy serializer for pytorch ( #12344 )" ( #12478 )
...
* [Core] zero-copy serializer for pytorch (#12344 )
* zero-copy serializer for pytorch
* address possible bottleneck
* add tests & device support
(cherry picked from commit 0a505ca83d )
* add environmental variables
* update doc
2020-11-30 11:43:03 -08:00
Sven Mika and GitHub
bb03e2499b
[RLlib] PyBullet Env native support via env str-specifier (if installed). ( #12209 )
2020-11-30 12:41:24 +01:00
Tao Wang and GitHub
b85c6abc3e
Rename fields/variables from client id to node id ( #12457 )
2020-11-30 14:33:36 +08:00
SangBin Cho and GitHub
3964defbe1
[Logging] Fix tensorflow logging issue. ( #12225 )
...
* in progress.
* ip
* In Progress
* done.
* fix lint.
* Addressed code review
* Addressed code review.
2020-11-29 22:16:52 -08:00
SangBin Cho and GitHub
91d54ef621
[Core] Remove actor arg from executor to allow users to specify actor… ( #12239 )
...
* [Core] Remove actor arg from executor to allow users to specify actor arg in their Actor.remote.
* Addressed code review.
2020-11-29 22:15:48 -08:00
chaokunyang and GitHub
17a6b9bbe7
Fix not cp jars ( #12456 )
2020-11-30 13:53:09 +08:00
Philipp Moritz and GitHub
cf73ccddae
Allow more fields for object metadata ( #12484 )
2020-11-29 21:50:18 -08:00
f1cc33a6a6
Actor resource backlog hotfix ( #12471 )
...
* prepare implemented
* works?
* deflek
* git
* deflek round 2
* .
* improve the test
Co-authored-by: Alex <alex@anyscale.com >
Co-authored-by: Eric Liang <ekhliang@gmail.com >
2020-11-29 20:55:50 -08:00
Sven Mika and GitHub
fb318addcb
[RLlib] Curiosity exploration module: tf/tf2.x/tf-eager support. ( #11945 )
2020-11-29 12:31:24 +01:00
Micah Yong and GitHub
a537b852e6
[docs][core] Documentation improvement in master/walkthrough.html ( #12473 )
2020-11-28 20:36:01 -08:00
8a406e1f9a
[SGD] Add PTL Docs ( #12440 )
...
Co-authored-by: Richard Liaw <rliaw@berkeley.edu >
2020-11-28 10:09:38 -08:00
60a545ab57
[RLLib] Fix HyperOptSearch tuple to list conversion ( #12462 )
...
Co-authored-by: Sumanth Ratna <sumanthratna@gmail.com >
2020-11-28 10:07:54 -08:00
Kai Fricke and GitHub
1d0ade1b93
Revert "[Core] zero-copy serializer for pytorch ( #12344 )" ( #12469 )
...
This reverts commit 0a505ca8
2020-11-28 10:06:02 -08:00
Eric Liang and GitHub
9ad0f173d6
Prestart workers to avoid slow start when multi-tenancy is enabled ( #12430 )
2020-11-27 21:47:46 -08:00
Sven Mika and GitHub
0df55a139c
[RLlib] Attention Net prep PR #1 : Smaller cleanups. ( #12447 )
...
* WIP.
* Fix.
* Fix.
* Fix.
2020-11-27 16:25:47 -08:00
Eric Liang and GitHub
569eee5e71
Enable more new scheduler tests ( #12421 )
2020-11-27 16:10:38 -08:00
Richard Liaw and GitHub
affb0b776c
Fix github issue template ( #12464 )
...
Signed-off-by: Richard Liaw <rliaw@berkeley.edu >
2020-11-27 14:13:29 -08:00
7c009d22cf
[docs] Add xgboost_ray to docs ( #12184 )
...
Co-authored-by: Amog Kamsetty <amogkamsetty@yahoo.com >
2020-11-27 11:36:56 -08:00
Siyuan (Ryans) Zhuang and GitHub
0a505ca83d
[Core] zero-copy serializer for pytorch ( #12344 )
...
* zero-copy serializer for pytorch
* address possible bottleneck
* add tests & device support
2020-11-26 16:09:54 -08:00
Amog Kamsetty and GitHub
e0573df337
[CI] Fix windows build ( #12415 )
...
* attempt to fix windows
* fix syntax
* try again
* try again
* try again
* Revert "[ray_client] Support calling functions from other functions and correct the tests (#12141 )"
This reverts commit 4066056a0d .
* Revert
* Revert "Revert "[ray_client] Support calling functions from other functions and correct the tests (#12141 )""
This reverts commit bb27b87b6c8d780ad796f4d4aeaa20113c8eca79.
* please work
* works
* fix
2020-11-26 10:52:11 -08:00
Sven Mika and GitHub
c1d7826bb7
[RLlib] Move pettingzoo from requirements.txt into requirements_rllib.txt ( #12400 )
2020-11-26 19:30:35 +01:00
Sven Mika and GitHub
6475297bd3
[RLlib] Torch LR schedule not working. Fix and added test case. ( #12396 )
2020-11-26 13:14:11 +01:00
fangfengbin and GitHub
d5215745e4
[PlacementGroup] Introduce GcsResourceManager and avoid copying resources when scheduling placement groups ( #12253 )
2020-11-26 11:21:58 +08:00
Edward Oakes and GitHub
90d7863eb3
Document that ray install-nightly command doesn't work for 1.0.1.post1 and older ( #12429 )
2020-11-25 19:55:12 -06:00
Ameer Haj Ali and GitHub
9ccf5f6ccc
[ray client] add metadata and secure options to Worker. ( #12409 )
2020-11-25 17:48:13 -08:00
Richard Liaw and GitHub
751e13a41e
[docs] redirect to discourse ( #12427 )
2020-11-25 17:10:10 -08:00
Richard Liaw and GitHub
323941c745
[tune] fix pbt flakey test ( #12418 )
2020-11-25 16:58:37 -08:00
Eric Liang and GitHub
f6a5b733d5
Remove flaky object manager test that's no longer needed
2020-11-25 12:45:47 -08:00
Ian Rodney and GitHub
679492a235
[serve] Use Long Polling in Backend Worker ( #12093 )
2020-11-25 12:11:38 -08:00
Ian Rodney and GitHub
ca6c2b2442
[docker] Use cuDNN7, not 8 ( #12375 )
2020-11-25 12:06:53 -08:00
SangBin Cho and GitHub
753cda2f28
[Dashboard] Delete old dashboard ( #12144 )
...
* Delete old dashboard from repo.
* Delete old dashboard from repo. 2
2020-11-25 11:31:02 -08:00
Sven Mika and GitHub
b7dbbfbf41
[RLlib] Issue 11591: SAC loss does not use PR-weights in critic loss term. ( #12394 )
...
* WIP.
* Fix and LINT.
2020-11-25 11:28:46 -08:00
Sven Mika and GitHub
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 and GitHub
841d93d366
[RLlib] Issue 12233 shared tf layers example not really shared (only works for tf1.x, not tf2.x). ( #12399 )
2020-11-25 11:27:19 -08:00
Sven Mika and GitHub
95175a822f
[RLlib] Issue 11974: Traj view API next-action (shift=+1) not working. ( #12407 )
...
* WIP.
* Fix and LINT.
2020-11-25 11:26:29 -08:00
Max Fitton and GitHub
2e95552f0c
[Dashboard] Defensive change to make sure we do not iterate over "None" in the case that workers is not present in node physical stats for a given node ( #12358 )
2020-11-25 11:06:45 -08:00
dc55f6ba3a
skip gcs fault tolerance test for the time being when new scheduler is enabled ( #12393 )
...
Co-authored-by: senlin.zsl <senlin.zsl@antfin.com >
2020-11-25 10:40:47 -08:00
Gekho457 and GitHub
2b293832e7
[Docker][minor] Update path in base-deps Dockerfile ( #12391 )
...
* Update path in base-deps Dockerfile
* development Dockerfile
2020-11-25 10:30:10 -08:00
SangBin Cho and GitHub
2e4e285ef0
[Object Spilling] Fusion small objects ( #12087 )
2020-11-25 10:13:32 -08:00
karstenddwx and GitHub
09d5413f70
[RLlib] rollout batch, handle rewards that are None (unknown) in a multi-agent env ( #11858 ) ( #11911 )
2020-11-25 13:39:22 +01:00
c009c178f6
[RLlib] Closes #11924 : Add support for custom/ray environments in rollouts.py for agents without workers ( #11926 )
...
* Closes #11924
Formerly, rollout.py would only load environments from gym (with
gym.make() ) , if an agent without workers is employed (such as ES or
ARS). This will result in an error, if a custom environment is used. This
PR adds the possibility to load environments from the ray registry,
while maintaining the support for gym environments.
* Update rllib/rollout.py
Co-authored-by: Sven Mika <sven@anyscale.io >
2020-11-25 08:43:17 +01:00
Tomasz Wrona and GitHub
82852f0ed2
[RLlib] Add ResetOnExceptionWrapper with tests for unstable 3rd party envs ( #12353 )
2020-11-25 08:41:58 +01:00
Ian Rodney and GitHub
c5845c3a4e
[docker] Docker stop on each node ( #12357 )
2020-11-24 23:15:53 -08:00