benchmark/experiment_runner: save accelerator_model on error #6218
Conversation
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
This commit can be seen as a partial revert of "63455e0cd Unify the way in which result files are dumped (pytorch#6162)". In that commit we missed that `experiment_cfg` does not have the `accelerator_model` record. Thus, when a benchmark fails we do not include that record in the JSONL file, and therefore resuming a run doesn't work because the failing entry is not recognized (note that when checking whether to resume we compare the JSONL entry against `benchmark_experiment`, which does have `accelerator_model`). We could fix this two ways: (1) always save `benchmark_experiment`, not only on success, or (2) add `accelerator_model` to experiment_config. I've chosen to go with (1) since that's what we were doing before 63455e0.
mbzomowski
pushed a commit
to mbzomowski-test-org/xla
that referenced
this pull request
Jan 3, 2024
…#6218) This commit can be seen as a partial revert of "63455e0cd Unify the way in which result files are dumped (pytorch#6162)". In that commit we missed that `experiment_cfg` does not have the `accelerator_model` record. Thus, when a benchmark fails we do not include that record in the JSONL file, and therefore resuming a run doesn't work because the failing entry is not recognized (note that when checking whether to resume we compare the JSONL entry against `benchmark_experiment`, which does have `accelerator_model`). We could fix this two ways: (1) always save `benchmark_experiment`, not only on success, or (2) add `accelerator_model` to experiment_config. I've chosen to go with (1) since that's what we were doing before 63455e0.
golechwierowicz
pushed a commit
that referenced
this pull request
Jan 12, 2024
This commit can be seen as a partial revert of "63455e0cd Unify the way in which result files are dumped (#6162)". In that commit we missed that `experiment_cfg` does not have the `accelerator_model` record. Thus, when a benchmark fails we do not include that record in the JSONL file, and therefore resuming a run doesn't work because the failing entry is not recognized (note that when checking whether to resume we compare the JSONL entry against `benchmark_experiment`, which does have `accelerator_model`). We could fix this two ways: (1) always save `benchmark_experiment`, not only on success, or (2) add `accelerator_model` to experiment_config. I've chosen to go with (1) since that's what we were doing before 63455e0.
bhavya01
pushed a commit
that referenced
this pull request
Apr 22, 2024
This commit can be seen as a partial revert of "63455e0cd Unify the way in which result files are dumped (#6162)". In that commit we missed that `experiment_cfg` does not have the `accelerator_model` record. Thus, when a benchmark fails we do not include that record in the JSONL file, and therefore resuming a run doesn't work because the failing entry is not recognized (note that when checking whether to resume we compare the JSONL entry against `benchmark_experiment`, which does have `accelerator_model`). We could fix this two ways: (1) always save `benchmark_experiment`, not only on success, or (2) add `accelerator_model` to experiment_config. I've chosen to go with (1) since that's what we were doing before 63455e0.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
This commit can be seen as a partial revert of "63455e0cd Unify the way in which result files are dumped (#6162)". In that commit we missed that
experiment_cfgdoes not have theaccelerator_modelrecord. Thus, when a benchmark fails we do not include that record in the JSONL file, and therefore resuming a run doesn't work because the failing entry is not recognized (note that when checking whether to resume we compare the JSONL entry againstbenchmark_experiment, which does haveaccelerator_model).We could fix this two ways: (1) always save
benchmark_experiment, not only on success, or (2) addaccelerator_modelto experiment_config.I've chosen to go with (1) since that's what we were doing before 63455e0.