Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Updates for annealing swafix branch from olmo-core#120

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.

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

Already on GitHub?Sign in to your account

Open
undfined wants to merge15 commits intomain
base:main
Choose a base branch
Loading
fromundfined/swafix-core
Open
Changes from1 commit
Commits
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
PrevPrevious commit
NextNext commit
comment
  • Loading branch information
@undfined
undfined committedJun 27, 2025
commita1ac0dae0841ad3cc502018e41b7dc8e3d0b9f9f
16 changes: 9 additions & 7 deletionssrc/cookbook/model/builder.py
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -647,13 +647,15 @@ def build(self) -> ModelTrainConfig:

self.transformer_config = self.transformer_config.merge(dotlist=self.model_overrides)

self.transformer_config.block.attention.sliding_window = SlidingWindowAttentionConfig(
force_full_attention_on_first_layer=False,
force_full_attention_on_last_layer=True,
pattern=[4096, 4096, 4096, -1],
)
self.transformer_config.block.attention.use_flash = True
self.transformer_config.block.attention.use_head_qk_norm = True
# TODO(undfined): The hax once swafix is not an issue anymore
if self.model_identifier == "olmo2_7B_swafix":
self.transformer_config.block.attention.sliding_window = SlidingWindowAttentionConfig(
force_full_attention_on_first_layer=False,
force_full_attention_on_last_layer=True,
pattern=[4096, 4096, 4096, -1],
)
self.transformer_config.block.attention.use_flash = True
self.transformer_config.block.attention.use_head_qk_norm = True

return ModelTrainConfig(
init_seed=self.seed,
Expand Down
[8]ページ先頭
©2009-2025 Movatter.jp