orbax issues

1

Add a multihost build job for Orbax Checkpoint.

copybara-service[bot]

Update docstring for `train_steps` in `EmergencyCheckpointManagerBenchmarkConfig`.

1

Update docstring for `train_steps` in `EmergencyCheckpointManagerBenchmarkConfig`.

copybara-service[bot]

Refactor GCS Operations to Use TensorFlow's File System

Summary: This PR decouples Orbax from direct GCS API dependencies by delegating all filesystem operations to the underlying TensorFlow gfile implementation. This change improves testability and makes Orbax less susceptible...

deepikarajani24

Add additional support for TENSORSTORE_GCS_BACKEND environment variable

1

Add additional support for TENSORSTORE_GCS_BACKEND environment variable This allows the user to configure the Tensorstore GCS backend (i.e. with `gcs` for http or `gcs_grpc` for grpc) for additional places in...

copybara-service[bot]

Build jobs with a multiprocess checkpoint benchmark job.

1

Build jobs with a multiprocess checkpoint benchmark job.

copybara-service[bot]

orbax uses logging.info directly and not the supplied log function, hence impossible to suppress logging w/o silencing the whole INFO log

1

orbax uses logging.info directly. e.g.: checkpoint_manager.py:620, composite_checkpoint_handler.py:502, bstract_checkpointer.py:35, async_checkpointer.py:170, checkpoint_manager.py:484, checkpoint_manager.py:1701, checkpoint.py:186, checkpoint_manager.py:801, checkpoint_manager.py:1987, checkpoint_manager.py:1408, sync_checkpointer.py:439, etc and not the supplied log function, hence impossible to suppress logging w/o silencing...

qwer1304