Cacher is the code snippet organizer for pro developers

x = tensor([[1,2], [9,4], [6,78]])
>>> torch.max(x, dim=1)
(tensor([ 2,  9, 78]), tensor([1, 0, 1]))


>>> y = tensor([1,2])
>>> torch.stack([y,y], dim=0)
tensor([[1, 2],
        [1, 2]])
>>> torch.stack([y,y], dim=1)
tensor([[1, 1],
        [2, 2]])
        
        
        
torch.bmm
>>> batch1 = torch.randn(10, 3, 4)
>>> batch2 = torch.randn(10, 4, 5)
>>> res = torch.bmm(batch1, batch2)
>>> res.size()
torch.Size([10, 3, 5])


unsqueeze / squeeze
energy = torch.bmm(query.unsqueeze(1), key.transpose(1, 2)).squeeze(
            dim=1)
            
softmax
self.softmax(torch.Tensor([[1,1, 8], [2,2, 2]]), )
tensor([[0.0009, 0.0009, 0.9982],
        [0.3333, 0.3333, 0.3333]])
     
---cat---   
>>> x = torch.randn(2, 3)
>>> x
tensor([[ 0.6580, -1.0969, -0.4614],
        [-0.1034, -0.5790,  0.1497]])
>>> torch.cat((x, x, x), 0)
tensor([[ 0.6580, -1.0969, -0.4614],
        [-0.1034, -0.5790,  0.1497],
        [ 0.6580, -1.0969, -0.4614],
        [-0.1034, -0.5790,  0.1497],
        [ 0.6580, -1.0969, -0.4614],
        [-0.1034, -0.5790,  0.1497]])
>>> torch.cat((x, x, x), 1)
tensor([[ 0.6580, -1.0969, -0.4614,  0.6580, -1.0969, -0.4614,  0.6580,
         -1.0969, -0.4614],
        [-0.1034, -0.5790,  0.1497, -0.1034, -0.5790,  0.1497, -0.1034,
         -0.5790,  0.1497]])
         
---zip---
a = [1,2]
b = ['a','b']
for x in zip(a,b):
...     print(x)
(1, 'a')
(2, 'b')

-----max----
torch.max(torch.Tensor([[1,2], [3,4]]), dim=1)
>>
[2,4]
[1, 1]