Pass train_state to update_params
#790
Open
Conversation
This file contains 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
…(shallow) copy in submission_runner
|
MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅ |
| @@ -422,6 +423,7 @@ def update_params(workload: Workload, | |||
| batch: Dict[str, Tensor], | |||
| loss_type: LossType, | |||
| optimizer_state: OptimizerState, | |||
| train_state: Dict[str, Any], | |||
There was a problem hiding this comment.
Can we make this an optional kwarg and default it to None for backwards compatibility?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 pull-request adds
train_stateto the arguments ofupdate_params.This allows for example to implement a time-based LR scheduler (see #785).
Since
train_stateis a dictionary, we have to avoid that a submission modifies its content.The solution adopted here passes
train_stateby copy. We use a shallow copying, since all values intrain_stateare primitive types, and cannot be modified in-place. Although very fast, this copying operation would however be counted as part of the submission time.A possible alternative solution that would avoid this problem would be to pass
train_state['accumulated_submission_time']directly instead of passing the entire dictionary.