Add comments for default value (pytorch#2242)

Author: yunjey

torch/nn/modules/batchnorm.py

@@ -67,7 +67,7 @@ class BatchNorm1d(_BatchNorm):
  momentum: the value used for the running_mean and running_var
  computation. Default: 0.1
  affine: a boolean value that when set to true, gives the layer learnable
- affine parameters.
+ affine parameters. Default: True
 
  Shape:
  - Input: :math:`(N, C)` or :math:`(N, C, L)`
@@ -114,7 +114,7 @@ class BatchNorm2d(_BatchNorm):
  momentum: the value used for the running_mean and running_var
  computation. Default: 0.1
  affine: a boolean value that when set to true, gives the layer learnable
- affine parameters.
+ affine parameters. Default: True
 
  Shape:
  - Input: :math:`(N, C, H, W)`
@@ -161,7 +161,7 @@ class BatchNorm3d(_BatchNorm):
  momentum: the value used for the running_mean and running_var
  computation. Default: 0.1
  affine: a boolean value that when set to true, gives the layer learnable
- affine parameters.
+ affine parameters. Default: True
 
  Shape:
  - Input: :math:`(N, C, D, H, W)`

torch/nn/modules/instancenorm.py

@@ -58,7 +58,7 @@ class InstanceNorm1d(_InstanceNorm):
  num_features: num_features from an expected input of size `batch_size x num_features x width`
  eps: a value added to the denominator for numerical stability. Default: 1e-5
  momentum: the value used for the running_mean and running_var computation. Default: 0.1
- affine: a boolean value that when set to true, gives the layer learnable affine parameters.
+ affine: a boolean value that when set to true, gives the layer learnable affine parameters. Default: False
 
  Shape:
  - Input: :math:`(N, C, L)`
@@ -102,7 +102,7 @@ class InstanceNorm2d(_InstanceNorm):
  num_features: num_features from an expected input of size batch_size x num_features x height x width
  eps: a value added to the denominator for numerical stability. Default: 1e-5
  momentum: the value used for the running_mean and running_var computation. Default: 0.1
- affine: a boolean value that when set to true, gives the layer learnable affine parameters.
+ affine: a boolean value that when set to true, gives the layer learnable affine parameters. Default: False
 
  Shape:
  - Input: :math:`(N, C, H, W)`
@@ -147,7 +147,7 @@ class InstanceNorm3d(_InstanceNorm):
  num_features: num_features from an expected input of size batch_size x num_features x depth x height x width
  eps: a value added to the denominator for numerical stability. Default: 1e-5
  momentum: the value used for the running_mean and running_var computation. Default: 0.1
- affine: a boolean value that when set to true, gives the layer learnable affine parameters.
+ affine: a boolean value that when set to true, gives the layer learnable affine parameters. Default: False
 
  Shape:
  - Input: :math:`(N, C, D, H, W)`