This repository has been archived by the owner on Nov 3, 2023. It is now read-only.
Fix RAG generation for T5 FiD #3657
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
I noticed a big discrepency between exact_match and token_em after I switched to using T5 FiD in public ParlAI. Per discussion with Kurt, it's cause the generate for T5 is supposed to call TorchGeneratorAgent, rather than using the T5 generation via Hugging Face. This was in the original itnernal implementation, but got lost at some point while open sourcing. Verified by printing out forced decoding outputs and the output from eval_model. Verified that when `token_em == 1` that the output of eval_model was also the same.
klshuster
approved these changes
May 19, 2021
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
I noticed a big discrepancy between exact_match and token_em after I switched to using T5 FiD in public ParlAI.
Per discussion with Kurt, it's cause the generate for T5 is supposed to call TorchGeneratorAgent, rather than using the T5 generation via Hugging Face. This was in the original itnernal implementation, but got lost at some point while open sourcing.
Verified by printing out forced decoding outputs and the output from eval_model. Verified that when
token_em == 1that the output of eval_model was also the same. Models trained with this also doing a lot better.