[RLlib] Minor rllib.utils cleanup.
#8932
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Can one of the admins verify this patch? |
|
Test PASSed. |
|
Test PASSed. |
sven1977
added
the
tests-ok
ericl
approved these changes
Jun 15, 2020
vonHartz
added a commit
to vonHartz/mdp-playground
that referenced
this pull request
Apr 30, 2021
RaghuSpaceRajan
added a commit
to automl/mdp-playground
that referenced
this pull request
Jun 4, 2021
* 1. plots grouped by agents/metrics, 2. code refactoring * code refactoring, documentation & plots cosmetic changes * made changes to display 4 plots per row * clear jupyter notebook plot output * clear jupyter notebook plot output * single launch file and dqn grid hp search * spider (radar) plot changes * analysis reshaping np array before plotting * solving merge conflicts * updated code to plot multiple meta features * mean over self.seed_idx to plotting methods instead of -1. * clear the plots * clear plots * fault tolerance & subsampling code * config.py markdown file * latex code generation * updated latex gen code * pulled changed from master and single_launcher branches (1) * improved fault tolerance check code * fixed auc error in radar plot * checkpoint changes * Added a finite horizon=100 env registered in Gym. * Undo ICML changes * Made package a PyPI package: installable by pip * Removed some problems from merge of Suresh and Jessica's branches. Added functionality to load data when not using a grid of configs. * fixed reqs, removed unused seed function, see ray-project/ray#8932 * udated for new ray version * moving toward pep8 compatability * some more pep8 * Added loading of stats when using random search agent configs instead of grid search. * Jupyter notebook that should have been in last commit: added Hydra analysis. * Added tests for run_experiments.py and reverted it to pre-merge code where it only runs Ray * # 310 Hydra configs on qbert: test run * # 310 actual * MAJOR: Added config_processor.py to move config processing to 1 place eventually. * Removed bug and small changes * config_processor: Moved most hacks out of run_experiments.py * Small changes * pikcle the tune analysis file * added docs * Set theme jekyll-theme-slate * removed old imports * added framework option lost in merge * fixes for ray 1.3.0 * Undo lost #TODOs * added log level command line arg * Small changes * MAJOR: Added random and sobol generation for configs; updated requirements in setup.py; undid Ray 1.3.0 compat. changes * # 311 Rainbow 100 random configs on qbert * # 311 Minor bug removed * # 311 actual * MAJOR: Added column mem_used_mb to CSV stats files. * # 312 DQN R noises with varying lrs * # 312 actual: Rolled back writing mem_used_mb to CSV stats to be compatible with old plotting code. * Transferred ownership to automl and updated bibtex. * Small changes * # 312 actual * testing poetry * playing with toml * mutex problem * fixed mutex extras * version bump for pypi test * added status badges * rebuild docs * rebuild docs * # 313 DQN P noises with varying lrs * # 313 actual * added build docs to git for subtree use * updated doc * added nojekyll * fixed command * fixed cov link * added workflows from master * # 296: with clip_rewards=False * added missing build command * # 314: test run * bumped versions for deploy test, moved tests * fixed typo * mujoco installation * debug mujoco install * debug mujoco install * Removed bug with GymEnvWrapper * # 315 DQN R noises with clip_rewards=False * # 314 actual * # 296 actual * added pypi isntall instructions, entry point (NOT FUNCTIONAL YET) * skip test, publish only when creating release * added author * rm mujoco installation * MAJOR: Added image_representations for cont. envs, removed some bugs and minor improvements * created main function * ensured prev_reward is always defined * another test commit to testpypi for testing installation via pip * reset build and release condition * renamed test workflow, added pypi installation test flow * removed ununsed packages" * ray installation fixed * added entry point to pyproject for poetry * version bump, deploy for pypi test * reset trigger condition * typo * generated lock file * added init to find run_exp * reset trigger condition * synced setup: updated urls, copied packages from poetry build * moved scripts around to fix entry point * version bump for pypi test * reset trigger condition * added cov to worklfow, fixed test for new script dir * allow to call main with args in other python scripts * Removed bug, small changes * MAJOR: Added grid environments * moved scripts to top level, adapted setups * Revert "moved scripts to top level, adapted setups" This reverts commit 19e18d8. * removed redundant import * reapplied cli change * reapplied cli entry point * set version for pypi * updated workflows for release prp, fixed old arg, made comment more specific * unified ray init that is version agnostic * updated readmes * removed other scripts for test * deps defined in setup.py, so use this to install * need np for configspace * need np for configspace * fixed gcc version, updated scipy version, added numpy to build reqs" * install setuptools, not np * added setup requirement * need manual numpy install * moved to toml * auto-install doesnt work * added ray extras" * used ray 0.9 for test due to beta_annealing_factor being depreceated * dont use mujoco in tests * manual tf install, fixed tmp_dir arg for ray 0.x * updated authors * rebuild docs * gym version * ray versions for test * updated ray version * typo * updated dep for test * install atari * Added grid env example and improved example.py * Changed grid envs to only allow motion in 1 dimension at a time. * Fixes for failing tests. * Added experimental branch to workflow tests * temporarily delete cluster script Co-authored-by: suresh-guttikonda <guttikondasuresh7@gmail.com> Co-authored-by: Jessica Borja <jessicalizethborja@hotmail.com> Co-authored-by: Jan Ole von Hartz <hartzj@cs.uni-freiburg.de>
vonHartz
added a commit
to automl/mdp-playground
that referenced
this pull request
Jun 7, 2021
* 1. plots grouped by agents/metrics, 2. code refactoring * code refactoring, documentation & plots cosmetic changes * made changes to display 4 plots per row * clear jupyter notebook plot output * clear jupyter notebook plot output * single launch file and dqn grid hp search * spider (radar) plot changes * analysis reshaping np array before plotting * solving merge conflicts * updated code to plot multiple meta features * mean over self.seed_idx to plotting methods instead of -1. * clear the plots * clear plots * fault tolerance & subsampling code * config.py markdown file * latex code generation * updated latex gen code * pulled changed from master and single_launcher branches (1) * improved fault tolerance check code * fixed auc error in radar plot * checkpoint changes * Added a finite horizon=100 env registered in Gym. * Undo ICML changes * Made package a PyPI package: installable by pip * Removed some problems from merge of Suresh and Jessica's branches. Added functionality to load data when not using a grid of configs. * fixed reqs, removed unused seed function, see ray-project/ray#8932 * udated for new ray version * moving toward pep8 compatability * some more pep8 * Added loading of stats when using random search agent configs instead of grid search. * Jupyter notebook that should have been in last commit: added Hydra analysis. * Added tests for run_experiments.py and reverted it to pre-merge code where it only runs Ray * # 310 Hydra configs on qbert: test run * # 310 actual * MAJOR: Added config_processor.py to move config processing to 1 place eventually. * Removed bug and small changes * config_processor: Moved most hacks out of run_experiments.py * Small changes * pikcle the tune analysis file * added docs * Set theme jekyll-theme-slate * removed old imports * added framework option lost in merge * fixes for ray 1.3.0 * Undo lost #TODOs * added log level command line arg * Small changes * MAJOR: Added random and sobol generation for configs; updated requirements in setup.py; undid Ray 1.3.0 compat. changes * # 311 Rainbow 100 random configs on qbert * # 311 Minor bug removed * # 311 actual * MAJOR: Added column mem_used_mb to CSV stats files. * # 312 DQN R noises with varying lrs * # 312 actual: Rolled back writing mem_used_mb to CSV stats to be compatible with old plotting code. * Transferred ownership to automl and updated bibtex. * Small changes * # 312 actual * testing poetry * playing with toml * mutex problem * fixed mutex extras * version bump for pypi test * added status badges * rebuild docs * rebuild docs * # 313 DQN P noises with varying lrs * # 313 actual * added build docs to git for subtree use * updated doc * added nojekyll * fixed command * fixed cov link * added workflows from master * # 296: with clip_rewards=False * added missing build command * # 314: test run * bumped versions for deploy test, moved tests * fixed typo * mujoco installation * debug mujoco install * debug mujoco install * Removed bug with GymEnvWrapper * # 315 DQN R noises with clip_rewards=False * # 314 actual * # 296 actual * added pypi isntall instructions, entry point (NOT FUNCTIONAL YET) * skip test, publish only when creating release * added author * rm mujoco installation * MAJOR: Added image_representations for cont. envs, removed some bugs and minor improvements * created main function * ensured prev_reward is always defined * another test commit to testpypi for testing installation via pip * reset build and release condition * renamed test workflow, added pypi installation test flow * removed ununsed packages" * ray installation fixed * added entry point to pyproject for poetry * version bump, deploy for pypi test * reset trigger condition * typo * generated lock file * added init to find run_exp * reset trigger condition * synced setup: updated urls, copied packages from poetry build * moved scripts around to fix entry point * version bump for pypi test * reset trigger condition * added cov to worklfow, fixed test for new script dir * allow to call main with args in other python scripts * Removed bug, small changes * MAJOR: Added grid environments * moved scripts to top level, adapted setups * Revert "moved scripts to top level, adapted setups" This reverts commit 19e18d8. * removed redundant import * reapplied cli change * reapplied cli entry point * set version for pypi * updated workflows for release prp, fixed old arg, made comment more specific * unified ray init that is version agnostic * updated readmes * removed other scripts for test * deps defined in setup.py, so use this to install * need np for configspace * need np for configspace * fixed gcc version, updated scipy version, added numpy to build reqs" * install setuptools, not np * added setup requirement * need manual numpy install * moved to toml * auto-install doesnt work * added ray extras" * used ray 0.9 for test due to beta_annealing_factor being depreceated * dont use mujoco in tests * manual tf install, fixed tmp_dir arg for ray 0.x * updated authors * rebuild docs * gym version * ray versions for test * updated ray version * typo * updated dep for test * install atari * Added grid env example and improved example.py * Changed grid envs to only allow motion in 1 dimension at a time. * Fixes for failing tests. * Added experimental branch to workflow tests * temporarily delete cluster script Co-authored-by: suresh-guttikonda <guttikondasuresh7@gmail.com> Co-authored-by: Jessica Borja <jessicalizethborja@hotmail.com> Co-authored-by: Jan Ole von Hartz <hartzj@cs.uni-freiburg.de>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR moves the
explained_varianceops (tf and torch) into tf_ops and torch_ops, repectively and erasesutils.seed.py(not used by anything).Related issue number
Checks
scripts/format.shto lint the changes in this PR.