[CPU] Add per-token asym INT8 dynamic quantization support to QKV/MLP node #27001
+1,928
−403
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
github-actions
bot
added
category: CPU
|
Hi @dmitry-gorokhov , this is only a draft due to some unsolved design decision, like how to conditionally enable dynamic quantization, and maybe on per-layer basis. please take a look and comment, Thanks! |
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Outdated suggestions cannot be applied.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.
Details:
to use AMX-INT8 to boost performance of QKV/MLP layers in LLM, we need dynamic per-token INT8 quantization according to many research papers (SmoothQuant for example) and reference implementations (xFT for example), here we add the support, when following conditions are met:
add AMX-FP16 support to QKV/MLP layers on GNR platform.
optimize single batch 2nd token special case.
Tickets: