v4.2.6

Bug Fixes and Other Changes

• Enable PT XLA distributed training on homogeneous clusters

v4.2.5

Bug Fixes and Other Changes

• relax exception type

v4.2.4

prepare release v4.2.4

v4.2.3

Bug Fixes and Other Changes

• update num_processes_per_host for smdataparallel runner

v4.2.2

Bug Fixes and Other Changes

• Removed version hardcoding for sagemaker test dependency
• update distribution_instance_group for pytorch ddp
• specify flake8 config explicitly

v4.2.1

Bug Fixes and Other Changes

• handle utf-8 decoding exceptions while processing stdout and stderr streams

v4.2.0

Features

• Heterogeneous cluster changes

v4.1.6

Bug Fixes and Other Changes

• update: protobuf version to overlap with TF requirements

v4.1.5

Bug Fixes and Other Changes

• Fix none exception class issue for mpi

v4.1.4

Bug Fixes and Other Changes

• Use framework provided error class and stack trace as error message