github
/
ColossalAI
mirror of https://github.com/hpcaitech/ColossalAI
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

[NFC] polish colossalai/nn/layer/parallel_3d/layers.py code style (#966)

pull/997/head
LuGY 2022-05-16 10:18:49 +08:00 committed by binmakeswell
parent 955463e542
commit fb5bc6cb28
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
colossalai/nn/layer/parallel_3d/layers.py
View File

@ -53,8 +53,8 @@ class LayerNorm3D(ParallelLayer):
self.weight = Parameter( self.weight = Parameter(
torch.ones(self.normalized_shape_per_partition, device=get_current_device(), dtype=dtype)) torch.ones(self.normalized_shape_per_partition, device=get_current_device(), dtype=dtype))
if bias: if bias:
self.bias = Parameter(torch.zeros(self.normalized_shape_per_partition, self.bias = Parameter(
device=get_current_device(), dtype=dtype)) torch.zeros(self.normalized_shape_per_partition, device=get_current_device(), dtype=dtype))
else: else:
self.bias = None self.bias = None
self.variance_epsilon = eps self.variance_epsilon = eps
@ -854,7 +854,7 @@ class PatchEmbedding3D(ParallelLayer):
input_ = split_tensor_3d(input_, 0, self.input_parallel_mode) input_ = split_tensor_3d(input_, 0, self.input_parallel_mode)
output = F.conv2d(input_, self.weight, self.bias, stride=self.patch_size) output = F.conv2d(input_, self.weight, self.bias, stride=self.patch_size)
if self.flatten: if self.flatten:
output = output.flatten(2).transpose(1, 2) # BCHW -> BNC output = output.flatten(2).transpose(1, 2) # BCHW -> BNC
cls_token = self.cls_token.expand(output.shape[0], -1, -1) cls_token = self.cls_token.expand(output.shape[0], -1, -1)
output = torch.cat((cls_token, output), dim=1) output = torch.cat((cls_token, output), dim=1)