build: move flash-linear-attention back to optional-dependencies#1894
Open
build: move flash-linear-attention back to optional-dependencies#1894
Conversation
Move flash-linear-attention from the [dependency-groups].dev group (PEP 735, resolvable only by this repo's own `uv sync`) into a dedicated `fla` [project.optional-dependencies] extra, and compose it into the `moe` and `all` extras. Downstream consumers (e.g. NeMo-RL) pull nemo-automodel via the `pkg[extra]` syntax, which cannot reference PEP 735 dependency-groups, so after #1580 `fla` was silently missing for them. Also drop the VCS SHA pin in favor of the official 0.4.2 release, now that `fla.ops.cp` (required for CP on linear attention) is included upstream. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> Signed-off-by: Zhaopeng Qiu <alexq@nvidia.com>
Signed-off-by: NeMo Bot <nemo-bot@nvidia.com>
Contributor
Author
|
/ok to test c27163b |
Contributor
Author
|
CI failures on L0_Unit_Tests_CPU / L0_Unit_Tests_GPU are pre-existing on main. This PR only touches pyproject.toml and is unrelated to the broken tests. |
Contributor
Contributor
|
/ok to test 046aa02 |
akoumpa
approved these changes
Apr 18, 2026
Contributor
|
@NVIDIA-NeMo/automation PTAL, thank you. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What does this PR do ?
Move
flash-linear-attentionfrom Automodel's[dependency-groups].dev(PEP 735) back into a dedicatedfla[project.optional-dependencies]extra so that downstream consumers (e.g. NeMo-RL) can resolve it transitively vianemo-automodel[moe]/[all]— PEP 735 dependency-groups cannot be referenced via thepkg[extra]syntax, which silently broke fla for consumers after #1580.Changelog
fla[project.optional-dependencies]extra pinningflash-linear-attention>=0.4.2(dropping the VCS SHA pin now that the official 0.4.2 release includesfla.ops.cp).nemo_automodel[fla]into themoeandallextras so downstream consumers pick fla up transitively.flash-linear-attention(and the stale "pin to main until a release includes it" comment) from[dependency-groups].dev.uv.lockto reflect the new resolution.Before your PR is "Ready for review"
Pre checks: