Conversation
Contributor
Author
|
fyi @tianyu-l @rakkit this will unblock deleting the token group padding logic from torchtitan (for everything including mxfp8) To clarify, the torchao So in Torchtitan, for
To:
Then when torch._grouped_mm executes and dispatches to torchao, we pad the groups, and unpad the outputs. |
tianyu-l
approved these changes
Mar 8, 2026
|
|
||
| filter_fqns: list[str] | ||
| mx_config: Any # MXLinearConfig type when imported | ||
| class MXFP8Converter(Configurable): |
Contributor
There was a problem hiding this comment.
Inherit QuantizationConverter
weifengpy
pushed a commit
to weifengpy/torchtitan
that referenced
this pull request
Mar 27, 2026
…torchao (pytorch#2520) ## Summary - Refactor MXFP8 model converter - Previous: 2 separate converters (linear, grouped_mm) - New: 1 unified converter for linear and grouped_mm ops - Details: pytorch/ao#3968 - Add `pad_token_groups_for_grouped_mm` config option to use dynamic per group padding kernels for MXFP8 grouped mm in torchao, so we can delete padding code from torchtitan (context: pytorch#2255) - torchao PR stack (must land first): pytorch/ao#4021 ## Tests - TODO: manually test this change prior to landing and update PR
TXacs
pushed a commit
to McmillanTAC/torchtitan
that referenced
this pull request
Apr 13, 2026
…torchao (pytorch#2520) ## Summary - Refactor MXFP8 model converter - Previous: 2 separate converters (linear, grouped_mm) - New: 1 unified converter for linear and grouped_mm ops - Details: pytorch/ao#3968 - Add `pad_token_groups_for_grouped_mm` config option to use dynamic per group padding kernels for MXFP8 grouped mm in torchao, so we can delete padding code from torchtitan (context: pytorch#2255) - torchao PR stack (must land first): pytorch/ao#4021 ## Tests - TODO: manually test this change prior to landing and update PR
ACharacterInASimulation
pushed a commit
to ACharacterInASimulation/torchtitan
that referenced
this pull request
Apr 21, 2026
…torchao (pytorch#2520) ## Summary - Refactor MXFP8 model converter - Previous: 2 separate converters (linear, grouped_mm) - New: 1 unified converter for linear and grouped_mm ops - Details: pytorch/ao#3968 - Add `pad_token_groups_for_grouped_mm` config option to use dynamic per group padding kernels for MXFP8 grouped mm in torchao, so we can delete padding code from torchtitan (context: pytorch#2255) - torchao PR stack (must land first): pytorch/ao#4021 ## Tests - TODO: manually test this change prior to landing and update PR
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
2 participants
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.
Summary
pad_token_groups_for_grouped_mmconfig option to use dynamic per group padding kernels for MXFP8 grouped mm in torchao, so we can delete padding code from torchtitan (context: Remove unnecessary token padding for MoE in BF16 mode #2255)Tests