Feature/sequence bucketing #610
Merged
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
…e cutting points for the different mini-batches
…thod also is checked if the object is not a list
|
@ArturoLlorente The IceTray disk space error is now fixed in main. Please try to update the branch and see if this doesn't fix the error. |
RasmusOrsoe
approved these changes
Oct 13, 2023
RasmusOrsoe
pushed a commit
to RasmusOrsoe/graphnet
that referenced
this pull request
Oct 25, 2023
…ence_bucketing Feature/sequence bucketing
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Outdated suggestions cannot be applied.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.
Implemented sequence bucketing as a collate function ( collator_sequence_bucketing) used by DataLoader module. This function sorts all graphs and splits the data into mini-batches, where the forward pass is performed separately.
Splits of the different mini-batches can be customized (by default 0-80%, 80-100%).
This method is effective when working with Transformers, since the computational complexity grows with the square of the sequence length. Improvement noted when then maximum number of pulses is greater.