Eric Liang
6724f57b03
[Examples] Add Carla test env ( #1343 )
...
* add carla example
* add reward
* set obs
* Sun Dec 17 16:06:00 PST 2017
* add spec
* fix measurement
* add train script
* resize to 80x80
* null
* initial small training run
* robustify env, clean up action space
* clean up vars
* switch to town2 which is faster
* tunify train.py
* add discrete mode
* update
* fix excessive brakinG
* fix the weather
* rename
* redirect output and from future import
* doc
* update
* fix rebase
* allow dqn gpu growht
* adjust dqn hyperparams
* better ppo parameters
2017-12-19 12:57:58 -08:00
Melih Elibol
24b93b1123
fixes default type for product of empty shape. ( #1341 )
2017-12-18 17:41:44 -08:00
Eric Liang
47b1f02d3e
[rllib] Pull out multi-gpu optimizer as a generic class ( #1313 )
2017-12-17 15:59:57 -08:00
Cathy Wu
53e736fe01
[rllib] Small fix for supporting custom preprocessors ( #1334 )
...
* Small fix for supporting custom preprocessors
* PEP8
* fix test
2017-12-17 04:37:29 -08:00
Eric Liang
bab44837e0
[tune] Tensorboard logger incorrectly reports training iteration as cur timestep value
2017-12-16 23:30:15 -08:00
Eric Liang
d21ea0ca45
Switch EC2 example config to use AWS deep learning AMI + latest Ray wheel ( #1331 )
...
* update
* install --user
2017-12-16 17:39:46 -08:00
Eric Liang
f5ea44338e
EC2 cluster setup scripts and initial version of auto-scaler ( #1311 )
2017-12-15 23:56:39 -08:00
Robert Nishihara
76b6b4a2d3
When killing worker, release resources before dispatching tasks. ( #1327 )
2017-12-15 18:12:03 -08:00
Eric Liang
fbf1806b8a
[tune] Clean up result logging: move out of /tmp, add timestamp ( #1297 )
2017-12-15 14:19:08 -08:00
Stephanie Wang
12fdb3f53a
Convert actor dummy objects to task execution edges. ( #1281 )
...
* Define execution dependencies flatbuffer and add to Redis commands
* Convert TaskSpec to TaskExecutionSpec
* Add execution dependencies to Python bindings
* Submitting actor tasks uses execution dependency API instead of dummy argument
* Fix dependency getters and some cleanup for fetching missing dependencies
* C++ convention
* Make TaskExecutionSpec a C++ class
* Convert local scheduler to use TaskExecutionSpec class
* Convert some pointers to references
* Finish conversion to TaskExecutionSpec class
* fix
* Fix
* Fix memory errors?
* Cast flatbuffers GetSize to size_t
* Fixes
* add more retries in global scheduler unit test
* fix linting and cast fbb.GetSize to size_t
* Style and doc
* Fix linting and simplify from_flatbuf.
2017-12-14 20:47:54 -08:00
Philipp Moritz
cac5f47600
First Part of Internal Ray API Refactor ( #1173 )
...
* add Ray status class
* add C++ util files
* add ID types
* more APIs
* build system integration
* add test infrastructure and implement some APIs
* add more tests
* fix bugs
* add task table tests
* update
* add toolchain file
* fix
* test
* link with pthread
* update
* fix
* more fixes
* fixes
* always vendor gtest and gflags
* linting
* fixes
* add constants file
* comments
* more fixes
* fix linting
2017-12-14 14:54:09 -08:00
Richard Liaw
c5c83a4465
[rllib] PPO and A3C unification ( #1253 )
2017-12-14 01:08:23 -08:00
Robert Nishihara
2f750e9ba7
Add parentheses around one-line if statement. ( #1318 )
2017-12-13 23:48:53 -08:00
Robert Nishihara
60d4f92d43
Add --user to instructions for building ray from source. ( #1319 )
2017-12-13 23:48:03 -08:00
Richard Liaw
cabbd27c56
[rllib] Support Nested Configuration Merging ( #1268 )
2017-12-13 14:39:01 -08:00
Robert Nishihara
f75b51d178
Register Common.error with local scheduler extension module. ( #1316 )
...
* Register Common.error with local scheduler extension module.
* Add test.
2017-12-13 11:55:54 -08:00
Richard Liaw
b6a35e0395
[rllib] Introduce pip install rllib ( #1310 )
...
* update setup
* more dependencies
2017-12-12 13:58:28 -08:00
Robert Nishihara
b1d89026cd
Make ActorMethod fields private to fix tab completion. ( #1312 )
2017-12-12 10:07:33 -08:00
Peter Schafhalter
20d6b74aa6
[rllib] Added evaluation script to RLLib ( #1295 )
2017-12-11 11:59:44 -08:00
Robert Nishihara
96c46d35ff
Tell Ray how to serialize FunctionSignature objects. ( #1308 )
2017-12-10 22:40:28 -08:00
Eric Liang
7009538321
Autodetect the number of GPUs when starting Ray. ( #1293 )
...
* autodetect
* Wed Dec 6 12:46:52 PST 2017
* Wed Dec 6 12:47:54 PST 2017
* Move GPU autodetection into services.py.
* Fix capitalization of Nvidia.
* Update documentation.
2017-12-09 15:30:16 -08:00
Robert Nishihara
6aae9a12fb
Improve version checking at startup. ( #1307 )
...
* Check pyarrow version at startup.
* For version check, use absolute path to ray module.
2017-12-09 14:20:56 -08:00
Robert Nishihara
96463c680c
Allow actor methods to return multiple object IDs. ( #1296 )
...
* Allow actor methods to return multiple object IDs.
* Add test.
* Fixes
* Remove outdated comment.
* Add comment and assert
2017-12-09 10:37:57 -08:00
Zongheng Yang
7e4a28f933
[rllib] Add tuned_examples/pong-ppo.yaml ( #1302 )
...
* Add tuned_examples/pong-ppo.yaml: 21 rew in ~3380sec
* Header comments
2017-12-09 01:20:22 -08:00
John Schulman
2606001a36
allow users to disable the webui ( #1306 )
...
* allow users to disable the webui
* Remove trailing whitespace.
2017-12-09 00:35:55 -08:00
Stephanie Wang
bac39a134e
Define a wrapper class for callback_data.data ( #1301 )
2017-12-08 11:48:21 -08:00
Robert Nishihara
5adbdfecd0
Raise exception if pyarrow is imported before ray. ( #1283 )
...
* Raise exception if pyarrow is imported before ray.
* Pip install pyarrow when building doc so we don't have to mock it.
* Raise ImportError instead of Exception.
2017-12-08 03:34:54 -08:00
Richard Liaw
2e0eb0e4c7
[rllib] Adding dependencies ( #1298 )
2017-12-08 01:57:19 -08:00
Philipp Moritz
26125e1547
Fixing the jenkins tests ( #1299 )
...
* trying to fix jenkins tests
* comment out more tests
* remove pytorch stuff
* use non-monotonic clock (monotonic not supported on python 2.7)
* whitespace
2017-12-07 17:03:58 -08:00
Eric Liang
35f7398666
[rllib] Update RLlib docs and README ( #1288 )
...
Updates the rllib docs and README.
2017-12-06 18:17:51 -08:00
Eric Liang
2d543b6e19
[rllib] Refactor DQN to use an Evaluator abstraction ( #1276 )
...
This introduces rllib.Evaluator and rllib.Optimizer classes. Optimizers encapsulate a particular distributed optimization strategy for RL. Evaluators encapsulate the model graph, and once implemented, any Optimizer may be "plugged in" to any algorithm that implements the Evaluator interface.
2017-12-06 17:51:57 -08:00
Stephanie Wang
044548bcff
Mark the killed as done outside of loop ( #1284 )
2017-12-02 14:42:16 -08:00
Robert Nishihara
c21e189371
Allow scheduling with arbitrary user-defined resource labels. ( #1236 )
...
* Enable scheduling with custom resource labels.
* Fix.
* Minor fixes and ref counting fix.
* Linting
* Use .data() instead of .c_str().
* Fix linting.
* Fix ResourcesTest.testGPUIDs test by waiting for workers to start up.
* Sleep in test so that all tasks are submitted before any completes.
2017-12-01 11:41:40 -08:00
Robert Nishihara
ac64631043
0.3 release blog post. ( #1274 )
...
* Initial draft of 0.3 release blog post.
* Small tweaks.
* Update date
* Updates based on comments
* Update date
* Update date.
* Fixes
2017-11-30 16:24:34 -08:00
Richard Liaw
483dee2ff3
[rllib] Generalizing A3C Sampling Classes ( #1250 )
2017-11-30 00:22:25 -08:00
Richard Liaw
7db07acc4f
website_analytics ( #1272 )
2017-11-29 10:01:59 -08:00
Robert Nishihara
dd45664ab5
Bump version number to 0.3.0. ( #1247 )
2017-11-27 23:02:29 -08:00
Eric Liang
37831ae0c3
Add a nicer warning message when you pass the wrong thing to ray.wait() ( #1239 )
...
* add warnings
* fix python mode
* Small changes and add tests.
* Fix test failure.
2017-11-27 22:57:33 -08:00
Robert Nishihara
c1496b8111
Check version info in ray start for non-head nodes. ( #1264 )
...
* Check version info in ray start for non-head nodes.
* Small fix.
* Fix
* Push error to all drivers when worker has version mismatch.
* Linting
* Linting
* Fix
* Unify methods.
* Fix bug.
2017-11-27 22:03:38 -08:00
Philipp Moritz
2c0d5544ac
Add streaming MapReduce example ( #1251 )
...
Add streaming MapReduce example.
2017-11-27 21:38:35 -08:00
Richard Liaw
5e37cb8e16
Small PPO bug ( #1265 )
2017-11-27 17:52:25 -08:00
Robert Nishihara
f7c4f41df8
Change Python Redis client psubscribe -> subscribe. ( #1261 )
2017-11-26 23:29:37 -08:00
Robert Nishihara
2865128df0
Remove counter from run_function_on_all_workers. Also remove utilitie… ( #1260 )
...
* Remove counter from run_function_on_all_workers. Also remove utilities for copying directories across machines.
* Fix linting.
2017-11-26 18:29:10 -08:00
Robert Nishihara
0b4961b161
Provide flag for setting redis maxclients. ( #1257 )
...
* Add flag for attempting to increase ulimit -n and the redis maxclients.
* Don't bother trying to set ulimit -n.
* Fix linting.
* Add basic test.
2017-11-26 18:25:55 -08:00
Eric Liang
7fc2ddbaf7
Revert "[rllib] Use NoFilter instead of MeanStdFilter for PPO. ( #1082 )" ( #1255 )
...
This reverts commit 971becc905
2017-11-26 16:00:46 -08:00
Robert Nishihara
e583d5a421
Give warnings for unimplemented Python mode methods. ( #1256 )
2017-11-26 13:11:12 -08:00
Eric Liang
81bed0fef8
[tune] Add internal implementation overview + image for task timeline ( #1254 )
2017-11-26 10:57:32 -08:00
Richard Liaw
f34d705178
[rllib] Update Docs for RLLib ( #1248 )
...
* init_changes
* last_changes
* addressing comments
* fix comments
* update
* nit
2017-11-24 10:36:57 -08:00
Robert Nishihara
7af5292646
Give error if a worker has a version mismatch for Python Ray, or clou… ( #1245 )
...
* Give error if a worker has a version mismatch for Python Ray, or cloudpickle.
* Check version when attaching driver to cluster.
* Only do check if the version info is present.
* Bug fix.
* Fix typo.
2017-11-23 23:31:03 -08:00
Eric Liang
ddfe00b7e8
[tune] Documentation for Ray.tune ( #1243 )
2017-11-23 11:31:59 -08:00
