This repository was archived by the owner on Apr 8, 2025. It is now read-only.
ELECTRA: use gelu for pooled output of ELECTRA model#364
Merged
brandenchan merged 1 commit intodeepset-ai:masterfrom May 14, 2020
Merged
Conversation
…al implementation
Contributor
Author
|
@brandenchan Would be interesting so see the performance diff :) |
Contributor
|
Great! I tested this on an Electra checkpoint and the performance is still good (in fact gelu got 0.2% averaged over 3 germeval tasks than tanh). Our CI is currently crashing but when I tested the branch locally, all tests passed! |
brandenchan
approved these changes
May 14, 2020
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 subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Hi,
this PR fixes the activation function for pooled output from the ELECTRA model, to match the original implementation.
geluis now used, wheras e.g. BERT usestanhas activation function. See discussion in #362.