Releases: AI-Hypercomputer/cloud-tpu-monitoring-debugging
Releases · AI-Hypercomputer/cloud-tpu-monitoring-debugging
v0.1.5
- Raise exception without waiting for daemon thread to terminate
Full Changelog: v0.1.4...v0.1.5
v0.1.4
- Gracefully exiting daemon threads
- Fixed the URL for PyPI package in README
Full Changelog: v0.1.3...v0.1.4
v0.1.3
- Fixing issue with using signals and threads together in a program
Full Changelog: v0.1.2...v0.1.3
v0.1.2
- Improved stack trace readability and clarity by adding a message for more information
Full Changelog: v0.1.1...v0.1.2
v0.1.1
- Bug Fixes
- Fixes dumping of stack traces on the console when exceptions like
AssertionError
,tensorflow.python.framework.errors_impl.NotFoundError
are thrown whencollect_stack_trace=True
andstack_trace_to_cloud=False
.
- Fixes dumping of stack traces on the console when exceptions like
- Updated README
Initial release of cloud-tpu-diagnostics PyPI package
What's Changed
- Initial release of cloud-tpu-diagnostics PyPI package that contains debug module to periodically collect stack traces on faults or when a program running on Cloud TPU is stuck or hung somewhere.
Full Changelog: https://github.com/google/cloud-tpu-monitoring-debugging/commits/v0.1.0