API for getting intermediate image and text features, forward_intermediates()#1035
Merged
API for getting intermediate image and text features, forward_intermediates()#1035
Conversation
…dd partial impl to CoCa model.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Compatible with timm approach (passes through to timm image encoders) and adding to builtin image and text...
This is a different approach than #731, but as with timm, I found the idea of integrating the output_hidden_states approach to be like HF Transformers to be a bit of a headache and risk too many regressions. This approach duplicates some code, but it makes typing less problematic, and keeps the main forward functions lean for training, normal use.