update master #2
Merged
Conversation
This file contains 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
* fixed bug in VecEnvWrapper.__getattr__ where inherited methods were inaccessible * improved test for VecEnvWrapper.__getattr__ to be more comprehensive * changed test function to satisfy code checks * updated changelog and simplified declaration of self.class_attributes * modified getattr_depth_check for consistency and added helper method for getting all attributes
* Fixed path splitting in _get_latest_run_id() on Windows machines * Returned to previous split method, replaced split delimiter with os.sep in _get_latest_run_id function. Wrote test for saving tensorboard data twice with the same logname * Fixed tests for saving tensorboard twice with same logname * Updated tensorboard tests * Updated tensorboard tests. Added name and fix to changelog * Update test_tensorboard.py * Update test_tensorboard.py
* added example project * format * some textual improvements * Update docs/misc/projects.rst Co-Authored-By: Antonin RAFFIN <[email protected]>
* Issue #317 [feature request] filter_size can be a array instead of one value * Issues #326 [Feature] filter_size can be a array * Issue #326 [Feature] filter_size can be a array * Issues #326 [Feature] filter_size can be a array: Line too long * Update changelog.rst * Issue #326 [Feature] filter_size can be a array, the added test code is test_a2c_conv.py * Issues #326 [Feature] filter_size can be a array, remove the unused variables * Issues #326 [Feature] filter_size can be a array, remove the unused library * Issue #326, [Feature] filter_size can be a array. Clean up the test code
* Minor typo correction * Minor typo correction
* fix timestep and learning rate recording * fix issue related to reset_num_timesteps * update changelog * Update changelog.rst
* Update doc: hyperparam tuning for rl zoo * Add colab notebook link
* * Add `get_parameters` function (returns all loadable/saveable tensorflow Variables)
* Add `load_parameters` function (loads model parameters from file/file-like/list of ndarrays)
* Update A2C, ACER, ACKTR, DDPG, DQN, PPOs, SAC and TRPO to use `get_parameters` to define
parameters necessary for correctly loading/saving models.
* * Switch from using lists of parameters to dicts of
variable name -> ndarray.
* Includes support for loading from older .pkl files
with a list of parameters
* Renamed `get_parameters` to `_get_parameter_list`
* `get_parameters` dictionary of variable name -> ndarrays
* `_get_parameter_list` returns list of tensorflow Variables
that should be saved/loaded
* Updated changelog for changes
* Clarified name of function parameter
* Updated contributor's list
* Fix few PEP8 errors
* Update docs to reflect variable name
* Fix PEP8/style in test_load_parameters
* Requested small typo/doc changes and removed unused parameter in tests
* Add parameter for with tests
* Add tests for from a file/file-like objects
* Use format-function and small line-length change
* Add warning about not updating trainer parameters upon 'load_parameters'
* Add an example of using load/get parameters with a simple ES example
* Use OrderedDict for get_params rather than normal dict
* Make _get_variable_list a public function
* Update load_parameters example with A2C+ES hybrid and only mutating specific parameters
* Add bit flipping env * HER reloaded (WIP) * DQN + HER * Add support for SAC and DDPG * Add tests for SAC and DDPG + HER + add comments * Bug fix + add comments * Add action noise for SAC * Add note about pop-art normalization * Add saving/loading + begin support for VecEnv * Add success rate * Fix HER learning method * Add support for VecEnv + improve comments + add properties to ReplayBuffer * Update documentation * Add HER example * Removed unused dependencies (tdqm, dill, progressbar2, seaborn, glob2, click) * Remove note on the replay buffer * Update doc + add a check for VecEnvWrapper with HER * Update examples + add notebook for HER * Add random exploration for SAC and DDPG * Typo in docstring * Doc update: add fix for DDPG saved models * Test with reward offset * Add GoalEnvNormalize draft * Remove GoalEnvNormalize * Fix typo * Bug fix for HER + VecEnv * Fix HER test env * Fixed key order * Add support for discrete obs space * Update doc about reproducing experiments * Update doc: DDPG supports multiprocessing with MPI * Fix for new abstract method * Update changelog * Fix custom policy example * Add replay_wrapper to base OffPolicy class * Fix reimport
* Bug fix when not enough samples in the replay buffer * Correct typo
* Remove buggy `find_trainable_variables` and replace it with `tf_util.get_trainable_vars` * Patch loading of old DDPG models * Fix indentation
* improved venenv doc * updated dummy_vec_env doc + improved vec_env doc * added VecCheckNan * added checking nan guide * added test * added hyperparam warning to doc * clean up and typos * codacy fixes + cleanup + changelog * hotfix * fix test * Update docs/guide/checking_nan.rst Co-Authored-By: Antonin RAFFIN <[email protected]> * Update docs/guide/checking_nan.rst Co-Authored-By: Antonin RAFFIN <[email protected]> * Update docs/guide/checking_nan.rst Co-Authored-By: Antonin RAFFIN <[email protected]> * Update docs/guide/checking_nan.rst Co-Authored-By: Antonin RAFFIN <[email protected]> * Update docs/guide/checking_nan.rst Co-Authored-By: Antonin RAFFIN <[email protected]> * Update docs/guide/checking_nan.rst Co-Authored-By: Antonin RAFFIN <[email protected]> * Update docs/guide/checking_nan.rst Co-Authored-By: Antonin RAFFIN <[email protected]> * fixed VecCheckNan exception only called once * add tf NaN debugging options to the NaN guide
* Clarify recommended SB usage + no tech support * Doc fix * Bump version * Update issue template * Update issue-template.md
* Fix GAE bug for TRPO/PPO1/GAIL * Update changelog.rst * Retrieve episode infos when using Monitor * Add comment about subsampling fisher vector product
…ning log probabilities (#397) * Support Gaussian probabilities and logp calculation * Fix linting + missing normalizer * Add & fix tests * Fix Gaussian PDF calculation * Bugfix in Gaussian probability calculation * Address review comments
* allow logging lists * Update docs/misc/changelog.rst Co-Authored-By: Antonin RAFFIN <[email protected]> * Move imports to top level * Add return type
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.
No description provided.