Add gigl.utils.tensorboard_writer.TensorBoardWriter#633
Draft
kmontemayor2-sc wants to merge 2 commits into
Draft
Add gigl.utils.tensorboard_writer.TensorBoardWriter#633kmontemayor2-sc wants to merge 2 commits into
gigl.utils.tensorboard_writer.TensorBoardWriter#633kmontemayor2-sc wants to merge 2 commits into
Conversation
…_metrics Introduces ``gigl.utils.tensorboard_writer.TensorBoardWriter``, the trainer/inferencer-side abstraction that writes scalar metrics to a Vertex AI ``ExperimentRun`` synchronously via ``aiplatform.log_time_series_metrics``. Key design points: - ``TensorBoardWriter.create(resource_name=..., experiment_name=..., experiment_run_name=..., enabled=is_chief_process)`` is the only constructor entry point. No env-var contract, no proto fields on ``GiglResourceConfig`` — configuration is plumbed through the trainer/inferencer's argparse (typically populated from ``GbmlConfig.trainerConfig.trainerArgs`` / ``inferencerConfig.inferencerArgs``). - ``enabled=False`` (non-chief ranks) returns a no-op writer. Chief ranks must supply all three string args; missing any of them raises ``RuntimeError`` so misconfiguration surfaces fast rather than producing a silent no-op. - Logs the cross-job experiment URL on ``start_run`` success so engineers can find the comparison TB page from trainer stdout. - Each ``log()`` is a single synchronous ``WriteTensorboardRunData`` RPC. Failures propagate to the caller (no background uploader thread). This PR introduces the writer with full test coverage but no callers; example trainer/inferencer entrypoints get wired up in a follow-up PR. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Collaborator
Author
|
/all_test |
Contributor
GiGL Automation@ 21:57:16UTC : 🔄 @ 21:59:18UTC : ✅ Workflow completed successfully. |
Contributor
GiGL Automation@ 21:57:16UTC : 🔄 @ 22:02:56UTC : ❌ Workflow failed. |
Contributor
GiGL Automation@ 21:57:17UTC : 🔄 @ 23:24:08UTC : ✅ Workflow completed successfully. |
Contributor
GiGL Automation@ 21:57:18UTC : 🔄 @ 23:01:32UTC : ✅ Workflow completed successfully. |
Contributor
GiGL Automation@ 21:57:19UTC : 🔄 @ 22:08:05UTC : ✅ Workflow completed successfully. |
Contributor
GiGL Automation@ 21:57:22UTC : 🔄 @ 23:33:02UTC : ✅ Workflow completed successfully. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Introduces
gigl.utils.tensorboard_writer.TensorBoardWriter, the trainer/inferencer-side abstraction that writes scalar metrics to a Vertex AIExperimentRunsynchronously viaaiplatform.log_time_series_metrics.Key design points:
TensorBoardWriter.create(resource_name=..., experiment_name=..., experiment_run_name=..., enabled=is_chief_process)is the only constructor entry point. No env-var contract, no proto fields onGiglResourceConfig— configuration is plumbed through the trainer/inferencer's argparse (typically populated fromGbmlConfig.trainerConfig.trainerArgs/inferencerConfig.inferencerArgs).enabled=False(non-chief ranks) returns a no-op writer. Chief ranks must supply all three string args; missing any of them raisesRuntimeErrorso misconfiguration surfaces fast rather than producing a silent no-op.start_runsuccess so engineers can find the comparison TB page from trainer stdout.log()is a single synchronousWriteTensorboardRunDataRPC. Failures propagate to the caller (no background uploader thread).This PR introduces the writer with full test coverage but no callers; example trainer/inferencer entrypoints get wired up in a follow-up PR.
Scope of work done
Where is the documentation for this feature?: N/A
Did you add automated tests or write a test plan?
Updated Changelog.md? NO
Ready for code review?: NO