diff options
author | soumith <soumith@fb.com> | 2016-12-12 13:49:06 +0300 |
---|---|---|
committer | soumith <soumith@fb.com> | 2016-12-12 13:49:06 +0300 |
commit | ee700e25cc47f0c7433d76059fb8bb72af62fb25 (patch) | |
tree | 2459c320875bfe070228833563860b29a2895e5d /doc | |
parent | 9b2f1ef7c45204f5c278bbdc55e367fcdd29e70e (diff) |
Support negative offset in nn.Narrownarrownegative
Diffstat (limited to 'doc')
-rw-r--r-- | doc/simple.md | 5 |
1 files changed, 2 insertions, 3 deletions
diff --git a/doc/simple.md b/doc/simple.md index b7044ae..09c60ca 100644 --- a/doc/simple.md +++ b/doc/simple.md @@ -25,7 +25,7 @@ Simple Modules are used for various tasks like adapting Tensor methods and provi * [MaskedSelect](#nn.MaskedSelect) : a [masked select](https://github.com/torch/torch7/blob/master/doc/tensor.md#tensor-maskedselect-index) module performs the torch.maskedSelect operation ; * [Index](#nn.Index) : a [index](https://github.com/torch/torch7/blob/master/doc/tensor.md#tensor-indexdim-index) over a given dimension ; * [Squeeze](#nn.Squeeze) : [squeezes](https://github.com/torch/torch7/blob/master/doc/tensor.md#tensor-squeezedim) the input; - * [Unsqueeze](#nn.Unsqueeze) : unsqueeze the input, i.e., insert singleton dimension; + * [Unsqueeze](#nn.Unsqueeze) : unsqueeze the input, i.e., insert singleton dimension; * [Transpose](#nn.Transpose) : [transposes](https://github.com/torch/torch7/blob/master/doc/tensor.md#tensor-transposedim1-dim2) the input ; * Modules that adapt mathematical Tensor methods : * [AddConstant](https://github.com/torch/nn/blob/master/doc/transfer.md#addconstant) : adding a constant ; @@ -679,8 +679,7 @@ The default is false. module = nn.Narrow(dimension, offset, length) ``` -Narrow is application of [narrow](https://github.com/torch/torch7/blob/master/doc/tensor.md#tensor-narrowdim-index-size) operation in a module. -The module further supports a negative `length` in order to handle inputs with an unknown size. +Narrow is application of [narrow](https://github.com/torch/torch7/blob/master/doc/tensor.md#tensor-narrowdim-index-size) operation in a module. The module further supports negative `length`, `dim` and `offset` to handle inputs of unknown size. ```lua > x = torch.rand(4, 5) |