diff options
| author | Sam Gross <sgross@fb.com> | 2016-02-25 01:18:13 +0300 |
|---|---|---|
| committer | Sam Gross <sgross@fb.com> | 2016-02-25 22:50:15 +0300 |
| commit | c327f5b86ad7972f12fd5fe41c7215eea381043d (patch) | |
| tree | 0edeaee6e9e069b069ae4feddd44dea60c1accaa /VolumetricBatchNormalization.lua | |
| parent | 00a60967687741ffdac20a8dda924127f8ccae4e (diff) | |
Add VolumetricBatchNormalization
The BatchNormalization modules now all extend nn.BatchNormalization and
use the same THNN/THCUNN implementation.
Diffstat (limited to 'VolumetricBatchNormalization.lua')
| -rw-r--r-- | VolumetricBatchNormalization.lua | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/VolumetricBatchNormalization.lua b/VolumetricBatchNormalization.lua new file mode 100644 index 0000000..6168a92 --- /dev/null +++ b/VolumetricBatchNormalization.lua @@ -0,0 +1,4 @@ +local BN, parent = torch.class('nn.VolumetricBatchNormalization', 'nn.BatchNormalization') + +-- expected dimension of input +BN.nDim = 5 |