Make EvoTorch future-proof #77
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
Some of the dependencies of EvoTorch became deprecated over time. This commit aims to replace those deprecated requirements with their up-to-date counterparts. In more details, the following changes are introduced: - If available, use `torch.func` instead of `functorch` - For getting the storage address of a tensor, use `x.untyped_storage().data_ptr()` instead of `x.storage().data_ptr()`, if available - Use `gymnasium` instead of `gym` - Introduce an EvoTorch-specific adapter between `gymnasium` and `brax`, instead of relying on the deprecated `gym` wrapper of `brax` Status: after the changes introduced by this commit, the unit tests pass successfully. Apart from the unit tests, test runs are still needed to ensure the correctness of these changes.
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## master #77 +/- ##
==========================================
- Coverage 77.83% 76.90% -0.94%
==========================================
Files 49 49
Lines 7332 7455 +123
==========================================
+ Hits 5707 5733 +26
- Misses 1625 1722 +97
☔ View full report in Codecov by Sentry. |
This commit updates the example notebooks demonstrating reinforcement learning and model predictive control, so that they adopt the API of the `gymnasium` library.
Brax v0.9.0 has a new set of implementations for its tasks. The old ones (which were the default until brax v0.1.2) are now within the namespace `brax.v1`. This commit extends `VecGymNE` with the ability to instantiate tasks from the namespace `brax.v1`, so that it will be possible to reproduce older results.
The pybullet humanoid example within the directory `examples/scripts/` is updated so that it works correctly with the new `GymNE` (which now uses the `gymnasium` library). At the moment of writing this, PyBullet environments (within the module `pybullet_envs`) do not have direct interfaces for `gymnasium`, and they fail with the latest version of the classical `gym` (i.e. gym v0.26). Therefore, for the pybullet humanoid example to work, one needs to install the classical `gym` like this: `pip install 'gym<0.26'` (while ensuring that `gymnasium` is also installed).
Higgcz
requested changes
Oct 2, 2023
for more information, see https://pre-commit.ci
Higgcz
approved these changes
Oct 2, 2023
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.
Some of the dependencies of EvoTorch became deprecated over time. This pull request aims to replace those deprecated requirements with their up-to-date counterparts. In more details, the following changes are introduced:
torch.funcinstead offunctorchx.untyped_storage().data_ptr()instead ofx.storage().data_ptr(), if availablegymnasiuminstead ofgymgymnasiumandbrax, instead of relying on the deprecatedgymwrapper ofbraxStatus.
Currently, the unit tests pass successfully. Apart from the unit tests, test runs are still needed to ensure the correctness of these changes.