Add T5 encoder exporting to MLIR and numerics verification with IREE #573
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
|
This PR is on top of #550 and needs it to be merged first. |
sogartar
force-pushed
the
t5-export
branch
2 times, most recently
from
d88edd4 to
c6f664f
Compare
sogartar
changed the title
KyleHerndon
approved these changes
Nov 20, 2024
This change adds the export functionality and a test for the T5 v1.1 encoder. Small and XXL variants. Added parsing of GGUF arrays when loading properties. This is used to deduce the vocabulary size instead of having to provide this manually. This can potentially be used also to load the tokenizer from GGUF or IRPA.
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Outdated suggestions cannot be applied.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 change adds the export functionality and a test for the T5 v1.1 encoder. Small and XXL variants.
The numerical tolerance in the test is
alto=1e-4 rtol=2.0e-3.I observed relative error of
1.47e-3on the XXL variant. This is probably OK. Definitely not outrageous.Added parsing of GGUF arrays when loading properties. This is used to deduce the vocabulary size instead of having to provide this manually. This can potentially be used also to load the tokenizer from GGUF or IRPA.