[LoRA] Continue optimizing MoE LoRA weight loading (#29322)

Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
This commit is contained in:
Jee Jee Li
2025-11-27 05:56:28 -08:00
committed by GitHub
parent cf348c8d27
commit 2f5f9acd55
15 changed files with 228 additions and 157 deletions
+2 -2
View File
@@ -87,7 +87,7 @@ class RowParallelLinearWithLoRA(BaseLinearLayerWithLoRA):
source_layer: nn.Module,
lora_config: LoRAConfig,
packed_modules_list: list,
model_config: PretrainedConfig | None,
model_config: PretrainedConfig | None = None,
) -> bool:
return type(source_layer) is RowParallelLinear
@@ -164,7 +164,7 @@ class RowParallelLinearWithShardedLoRA(RowParallelLinearWithLoRA):
source_layer: nn.Module,
lora_config: LoRAConfig,
packed_modules_list: list,
model_config: PretrainedConfig | None,
model_config: PretrainedConfig | None = None,
) -> bool:
# specifying kwargs so they can be easily accessed in decorator
return super().can_replace_layer(