{{ message }}
Repository navigation
Support int8 convrot embedding lookup (CORE-371) - #15035
Merged
Merged
Conversation
kijai
requested review from
Kosinkadink,
alexisrolland,
comfyanonymous,
guill and
rattus128
as code owners
July 22, 2026 17:27
comfyanonymous
approved these changes
Jul 28, 2026
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.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.

Adds support for int8 embedding lookup, useful for text generation models with lm_head or embed_tokens layers such as used in Gemma4.
Compared to the fp8 scaled (the only previously supported quant format for embed lookup), int8 with convrot reaches much lower relative error at ~0.8% vs ~2.6% of fp8_scaled.
Since the embedding lookup is only ~0.10% of the token time, speed is not an issue and I only included the eager backend.
Requires comfy-kitchen PR: Comfy-Org/comfy-kitchen#83