Update ml-goodput-measurement package for prod PyPi release v0.0.13.

PiperOrigin-RevId: 787142637

Author: dipannita08

CHANGELOG.md

@@ -20,6 +20,10 @@ To release a new version (e.g. from `1.0.0` -> `2.0.0`):
   * Update the `[Unreleased]` url: `v1.0.0...HEAD` -> `v2.0.0...HEAD`
 
 -->
+## [0.0.13] - 2025-07-25
+
+* Fix occasional gaps in cumulative Goodput Monitor dashboard.
+
 ## [0.0.12] - 2025-06-25
 
 * Support monitoring disruption badput due to infrastructure recovery.
@@ -97,6 +101,7 @@ To release a new version (e.g. from `1.0.0` -> `2.0.0`):
 * Initial release of ML Goodput Measurement PyPi package
 * Feature: Contains the Goodput module which allows logging and retrieval of training job's overall productive Goodput
 
+[0.0.13]: https://github.com/AI-Hypercomputer/ml-goodput-measurement/compare/v0.0.12...v0.0.13
 [0.0.12]: https://github.com/AI-Hypercomputer/ml-goodput-measurement/compare/v0.0.11...v0.0.12
 [0.0.11]: https://github.com/AI-Hypercomputer/ml-goodput-measurement/compare/v0.0.10...v0.0.11
 [0.0.10]: https://github.com/AI-Hypercomputer/ml-goodput-measurement/compare/v0.0.8...v0.0.10

README.md

@@ -410,6 +410,18 @@ class BadputType(enum.Enum):
   detection of a planned disruption signal. The save operation in type of
   checkpointing is synchronous resulting in time lost to Badput.
 
+  > **_NOTE:_** This type of Badput is reported *only* when Orbax is used for
+     checkpointing and requires the Orbax structured logger to be configured.
+     To compute checkpointing Badput for other types of checkpointers (Non-Orbax),
+     please use the Custom Badput Recorder API (instructions in
+     [Record Custom Badput Events (e.g., Evaluation, SDC Checks)](#record-custom-badput-events-eg-evaluation-sdc-checks))
+     with an appropriate Custom Badput event type and wrap the blocking operation
+     around the `start` and the `stop` API calls.
+
+  > **_NOTE:_** Do **NOT** use the Custom Badput APIs for blocking checkpoint save
+     operations if you are using Orbax. Either use Orbax's structured checkpoint
+     logger **OR** the Custom Badput API for any other type of checkpointing.
+
  - Wasted Progress due to Disruption (WASTED_PROGRESS_FROM_DISRUPTION)
 
  Based on checkpointing frequency, a disruption may result in time lost in the

pyproject.toml

@@ -14,7 +14,7 @@
 
 [project]
 name = "ml_goodput_measurement"
-version = "0.0.12"
+version = "0.0.13"
 authors = [
   { name="Cloud TPU Team", email="[email protected]" },
 ]