mirror of
https://github.com/deepseek-ai/DeepSeek-V3.git
synced 2025-04-19 01:59:01 -04:00
Merge pull request #556 from XxAlonexX/main
Fix Linear Layer Bias Initialization
This commit is contained in:
commit
87a01053e4
@ -185,7 +185,7 @@ class Linear(nn.Module):
|
|||||||
else:
|
else:
|
||||||
self.register_parameter("scale", None)
|
self.register_parameter("scale", None)
|
||||||
if bias:
|
if bias:
|
||||||
self.bias = nn.Parameter(torch.empty(self.part_out_features))
|
self.bias = nn.Parameter(torch.empty(out_features))
|
||||||
else:
|
else:
|
||||||
self.register_parameter("bias", None)
|
self.register_parameter("bias", None)
|
||||||
|
|
||||||
|
Loading…
Reference in New Issue
Block a user