This repository has been archived by the owner on Aug 7, 2024. It is now read-only.
add per-gemm config to Float8LinearConfig
#334
Closed
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
Summary: Previously the per-gemm configuration had to be hardcoded in library code. This PR exposes it to the top-level UX by adding a `Float8GemmConfig` field to `Float8LinearConfig`. Note that today the only supported configuration option is `use_fast_accum`. In the future, configuring output_dtype and whether to keep a gemm in higher precision would go here. Test Plan: ``` ./test/test_everything.sh ``` Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
This was referenced Jul 25, 2024
facebook-github-bot
added
the
CLA Signed
Summary: Previously the per-gemm configuration had to be hardcoded in library code. This PR exposes it to the top-level UX by adding a `Float8GemmConfig` field to `Float8LinearConfig`. Note that today the only supported configuration option is `use_fast_accum`. In the future, configuring output_dtype and whether to keep a gemm in higher precision would go here. Test Plan: ``` ./test/test_everything.sh ``` Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
vkuzo
commented
Jul 25, 2024
| ScaledMMConfig( | ||
| emulate, | ||
| True if not emulate else False, |
There was a problem hiding this comment.
this was mistakenly set to default True in #315, fixing
This was referenced Jul 25, 2024
weifengpy
approved these changes
Jul 25, 2024
|
@vkuzo has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
This pull request has been merged in b9b606e. |
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Outdated suggestions cannot be applied.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.
Stack from ghstack (oldest at bottom):
Float8LinearConfig#334DelayedScalingRecipetoDelayedScalingConfig#333config.enable_fsdp_fp8_all_gatherto usefloat8#332Summary:
Previously the per-gemm configuration had to be hardcoded in library
code. This PR exposes it to the top-level UX by adding a
Float8GemmConfigfield toFloat8LinearConfig.Note that today the only supported configuration option is
use_fast_accum. In the future, configuring output_dtypeand whether to keep a gemm in higher precision would go here.
Test Plan:
Reviewers:
Subscribers:
Tasks:
Tags:
Differential Revision: D60252069