Update vectorized reinforcement learning #104
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
This commit updates vectorized reinforcement functionalities of EvoTorch, so that they are compatible with the gymnasium 1.0.x API (while preserving compatibility with gymnasium 0.29.x). In more details, this commit introduces an EvoTorch-specific `SyncVectorEnv` implementation (as an alternative to gymnasium's SyncVectorEnv class). This custom SyncVectorEnv preserves the classical auto-reset behavior on which EvoTorch relies, allowing us to transition to gymnasium 1.0.x. Having our custom `SyncVectorEnv` allows us to introduce these performance-related improvements as well: (i) observations, rewards, etc. reported by `SyncVectorEnv` are now moved into the device of where the policies are executed; (ii) sub-environments of SyncVectorEnv that have reached the maximum number of episodes are not executed further. Brax-related notebook examples are also refactored. Instead of including the entire brax example in a single notebook, there are now two notebooks, one focusing on the training and the other focusing on the visualization. The visualization notebook is updated so that it works correctly with the latest version of brax.
engintoklu
requested review from
Higgcz,
pliskowski,
flukeskywalker and
NaturalGradient
Fix a typo in the code
flukeskywalker
approved these changes
Jun 6, 2024
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Outdated suggestions cannot be applied.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 pull request updates vectorized reinforcement functionalities of EvoTorch, so that they are compatible with the gymnasium 1.0.x API (while preserving compatibility with gymnasium 0.29.x).
In more details, this pull request introduces an EvoTorch-specific
SyncVectorEnvimplementation (as an alternative to gymnasium'sSyncVectorEnvclass). This custom SyncVectorEnv preserves the classical auto-reset behavior on which EvoTorch relies, allowing us to transition to gymnasium 1.0.x.Having our custom
SyncVectorEnvallows us to introduce these performance-related improvements as well:SyncVectorEnvare now moved into the device of where the policies are executed;Brax-related notebook examples are also refactored. Instead of including the entire brax example in a single notebook, there are now two notebooks, one focusing on the training and the other focusing on the visualization. The visualization example is updated so that it works correctly with the latest version of brax.