Releases: google/orbax
Releases · google/orbax
v0.1.5
Added
- Use a more precise timestamp when generating temporary directory names to
permit more than one concurrent checkpointing attempt per second. - Automatic import of nest_asyncio.
v0.1.4
Added
- Support for generic transformation function in PyTreeCheckpointHandler.
- Support n-digit checkpoint step format.
Fixed
- Eliminate Flax dependency to fix circular dependency problem.
v0.1.3
Added
sharding
option on `ArrayRestoreArgs
v0.1.2
Added
- Add "standard user recipe" to documentation.
- Add unit tests using mock to simulate preemption.
- Logging to increase transparency around why checkpoints are kept vs. deleted.
- Expand on uses of restore_args in colab.
- Expose utils_test.
- Add msgpack_utils to move toward eliminating Flax dependency.
- CheckpointManager starts a background thread to finalize checkpoints so that
checkpoints are finalized as soon as possible in async case.
Changed
- Remove CheckpointManager update API.
- Remove support for deprecated GDA.
- Add tmp suffix on step directory creation in CheckpointManager.save.
Fixed
- Preemption when using keep_time_interval caused the most recent steps before
preemption to be kept, despite not falling on the keep time interval.
v0.1.1
Added
- A util function that constructs restore_args from a target PyTree.
- CheckpointManager
delete
API, which allows deleting an existing step. - Made dev dependencies optional to minimize import overhead.
Changed
- Refactored higher-level utils in checkpoint_utils, which provides user-convenience functions.
- Guard option to create top-level directory behind
create
option. - Remove support for Python 3.7.
v0.1.0
Added
- Check for metric file in addition to item directory in CheckpointManager.
- Additional logs to indicate save/restore completion.
- Support for None leaves in PyTree save/restore.
- ArrayCheckpointHandler for individual arrays/scalars.
read: bool
option on all_steps to force read from storage location instead
of using cached steps.- Simplified "Getting Started" section in the docs.
- CheckpointManager creates the top level directory if it does not yet exist.
- Write msgpack bytes asynchronously.
Changed
- Removed some unused test_utils methods for filtering empty nodes.
- Update docs on
PyTreeCheckpointHandler
. - Removed unneeded AbstractCheckpointManager.
Fixed
- Usage of bytes_limiter to prevent too many bytes from being read during a
single restore call. - Temp checkpoint cleanup when using a step prefix (i.e. 'checkpoint_0').
v0.0.23
Added
- Option to customize metadata file name for Tensorstore.
Fixed
- Restore failure on GCS due to misidentification of checkpoint as
"not finalized".
v0.0.22
Added
- Added CHANGELOG.md for version updates (additions and changes), ingested by
auto-publish functionality.
v0.0.19
What's Changed
- [Orbax] Log unmatched keys from the new tree by @copybara-service in #147
- Fix auto-publish tool, replace setup.py with pyproject.toml standard, disable duplicate builds. by @copybara-service in #154
- Add support for tracking checkpoint metrics with Orbax in T5X. by @copybara-service in #153
Full Changelog: v0.0.18...v0.0.19
v0.0.18
What's Changed
- Add auto-publish tool for Orbax. by @copybara-service in #142
Full Changelog: v0.0.16...v0.0.18