Talk Is Cheap

1，下载并启动zbstudio$ git clone https://github.com/soumith/zbs-torch.git $ cd zbs-torch $ ./zbstudio.sh 2，安装Torch http://torch.ch/docs/getting-started.html3，安装ncclgit clone https://github.com/NVIDIA/n

th> a = torch.ones(5)*2th> a2 2 2 2 2th> b = torch.ones(5)*3th> b3 3 3 3 3th> nn.CMulTable():forward({a,b})6 6 6 6 6th> nn.CAddTable():forward({a,b})5 5 5 5 5

th> a = torch.ones(5,2)th> a1 1 1 1 1 1 1 1 1 1th> b = nn.View(2,-1)(a)th> b1 1 1 1 1 1 1 1 1 1th> a = torch.ones(12,2)th> b = nn.View(4,3,2)(a)th> b(1,.,.) = 1 1 1

th> a = torch.ones(3,2,2)th> b = nn.SpatialZeroPadding(0, 0, 1)(a)th> b(1,.,.) = 0 0 1 1 1 1(2,.,.) = 0 0 1 1 1 1(3,.,.) = 0 0 1 1 1 1th> b = nn.SpatialZeroPadd

th> a = torch.ones(4,3,2)th> b = nn.SpatialAveragePooling(2,2)(a)th> b(1,.,.) = 1 1(2,.,.) = 1 1(3,.,.) = 1 1(4,.,.) = 1 1th> a = torch.ones(3,2,2)th> b = nn.SpatialAvera

https://github.com/torch/nn/blob/master/doc/simple.md#nn.MMth> a = torch.ones(1,2,3)th> b = torch.ones(1,3,4)th> nn.MM()({a,b})(1,.,.) = 3 3 3 3 3 3 3 3th> a = torch.ones(1,2) th> b = t

https://github.com/torch/nn/tree/master/doc https://github.com/torch/torch7/tree/master/doc

Criterions有其forward和backward函数 https://github.com/torch/nn/blob/master/doc/criterion.md Module也有其forward和backward函数 https://github.com/torch/nn/blob/master/doc/module.mdModule的forward函数最简单，就是输入input