Skip to content

Bump flash attention 2 #3403

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
francesco-bertolotti wants to merge 1 commit into
base: main
Choose a base branch
from
+6 −4
Open

Bump flash attention 2 #3403

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 5 additions & 3 deletions tests/pytorch/attention/test_attention.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -104,8 +104,10 @@ def reset_global_fp8_state():
("2.8.3", True),
("2.8.3+local_version", True),
("2.8.3.post1", True),
("2.8.4", False),
("2.8.4+local_version", False),
("2.8.4", True),
("2.8.4+local_version", True),
("2.8.5", False),
("2.8.5+local_version", False),
Comment thread
greptile-apps[bot] marked this conversation as resolved.
("2.9.0", False),
),
)
Expand All @@ -125,7 +127,7 @@ def test_flash_attention_supported_version_message():
_get_supported_versions(
FlashAttentionUtils.version_required, FlashAttentionUtils.max_version
)
== ">= 2.1.1, < 2.8.4"
== ">= 2.1.1, < 2.8.5"
)


Expand Down
2 changes: 1 addition & 1 deletion transformer_engine/pytorch/attention/dot_product_attention/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -122,7 +122,7 @@ class FlashAttentionUtils:
version = PkgVersion("0")
version_required = PkgVersion("2.1.1")
version_required_blackwell = PkgVersion("2.7.3")
max_version = PkgVersion("2.8.4")
max_version = PkgVersion("2.8.5")
v2_plus = False
v2_1_plus = False
v2_3_plus = False
Expand Down