Actions: Ancientea/CannotMax
Actions
Showing runs from all workflows
105 workflow runs
105 workflow runs
Refactor UnitAwareTransformer to PhasePreservingLinearMHA
Build and Release
#91:
Commit 5494ed6
pushed
by
liemark
Refactor UnitAwareTransformer with gated attention
Build and Release
#90:
Commit f58dbd2
pushed
by
liemark
Reduce layers in UnitAwareTransformer and adjust logic
Build and Release
#89:
Commit 8fb997c
pushed
by
liemark
Clamp left and right features to prevent Inf values
Build and Release
#85:
Commit 8f8ed18
pushed
by
liemark
Add residual scaling for stability in forward pass
Build and Release
#83:
Commit a142d5d
pushed
by
liemark
Add residual scaling for stability in forward pass
Build and Release
#82:
Commit a142d5d
pushed
by
liemark
Update learning rate comments in train.py
Build and Release
#81:
Commit 252bb16
pushed
by
liemark
Replace fully connected layer with L2 norm calculations
Build and Release
#80:
Commit d8126b4
pushed
by
liemark
Remove normalization layers from model
Build and Release
#79:
Commit 6ccf39c
pushed
by
liemark
feat(ArknightsApp): 添加PC窗口连接和截图方式选择功能
Build and Release
#78:
Commit 91e4adb
pushed
by
AbyssRei
feat(ArknightsApp): 添加PC窗口连接和截图方式选择功能
Build and Release
#77:
Commit 91e4adb
pushed
by
AbyssRei
refactor(maa_adb_connector): 移除无用的maafw路径解析逻辑,简化代码结构
Build and Release
#76:
Commit 340099c
pushed
by
AbyssRei
Delete dark_mode_style_fix.py
Build and Release
#74:
Commit 48a4c33
pushed
by
liemark
Delete simular_history_match_ui.py
Build and Release
#71:
Commit 3d24710
pushed
by
liemark
Delete similar_history_match.py
Build and Release
#70:
Commit 6f1c9b5
pushed
by
liemark