Skip to content

Remove repeating flash_attention options.#2312

Merged
regisss merged 1 commit into
huggingface:mainfrom
ugolowic:flash_attn_arguments_fix
Oct 17, 2025
Merged

Remove repeating flash_attention options.#2312
regisss merged 1 commit into
huggingface:mainfrom
ugolowic:flash_attn_arguments_fix

Conversation

@ugolowic
Copy link
Copy Markdown
Collaborator

Some flash attention options were duplicated in training args and in run_lora_clm script which caused an ArgumentError.

This is a quick fix to remove the error. A broader unification of the arguments would be nice later.

Some flash attention options were duplicated in training args
and in run_lora_clm script which caused an ArgumentError.

This is a quick fix to remove the error. A broader unification
of the arguments would be nice later.

Signed-off-by: Urszula <urszula.golowicz@intel.com>
@ugolowic ugolowic requested a review from regisss as a code owner October 17, 2025 08:10
@HuggingFaceDocBuilderDev
Copy link
Copy Markdown

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

Copy link
Copy Markdown
Collaborator

@karol-brejna-i karol-brejna-i left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

Copy link
Copy Markdown
Collaborator

@regisss regisss left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@regisss regisss merged commit aae3792 into huggingface:main Oct 17, 2025
2 of 4 checks passed
ugolowic added a commit that referenced this pull request Oct 17, 2025
Signed-off-by: Urszula <urszula.golowicz@intel.com>
gplutop7 pushed a commit to HabanaAI/optimum-habana-fork that referenced this pull request Nov 6, 2025
…face#778)

Signed-off-by: Urszula <urszula.golowicz@intel.com>
Co-authored-by: Urszula Golowicz <urszula.golowicz@intel.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants