DEV Community

Super Kai (Kazuya Ito)
Super Kai (Kazuya Ito)

Posted on • Edited on

matmul and dot in PyTorch

Buy Me a Coffee

*My post explains mv(), mm() and bmm().

matmul() can do dot, matrix-vector or matrix multiplication with two of the 1D or more D tensors of zero or more elements, getting the 0D or more D tensor of one or more elements:

*Memos:

  • matmul() can be used with torch or a tensor.
  • The 1st argument(input) with torch or using a tensor(Required-Type:tensor of int, float or complex). *It must be a 1D or more D tensor.
  • The 2nd argument with torch or the 1st argument with a tensor is other(Required-Type:tensor of int, float or complex). *It must be a 1D or more D tensor.
  • There is out argument with torch(Optional-Defaual:None-Type:tensor): *Memos:
    • out= must be used.
    • My post explains out argument.
  • The combination of a 1D tensor(input or a tensor) and a 1D tensor(other) is done by dot multiplication.
  • The combination of a 2D or more D tensor(input or a tensor) and a 1D tensor(other) is done by matrix-vector multiplication.
import torch # Dot multiplication  tensor1 = torch.tensor([2, -5, 4]) tensor2 = torch.tensor([3, 6, -1]) torch.matmul(input=tensor1, other=tensor2) tensor1.matmul(other=tensor2) # tensor(-28)  # Matrix-vector multiplication  tensor1 = torch.tensor([[2, -5, 4], [-9, 0, 6]]) tensor2 = torch.tensor([3, 6, -1]) torch.matmul(input=tensor1, other=tensor2) # tensor([-28, -33])  # Matrix multiplication  tensor1 = torch.tensor([[2, -5, 4], [-9, 0, 6]]) tensor2 = torch.tensor([[3, 6, -1, 9], [-8, 0, 7, -2], [-7, -3, -4, 5]]) torch.matmul(input=tensor1, other=tensor2) # tensor([[18, 0, -53, 48], # [-69, -72, -15, -51]])  tensor1 = torch.tensor([2, -5, 4]) tensor2 = torch.tensor([[3, 6, -1, 9], [-8, 0, 7, -2], [-7, -3, -4, 5]]) torch.matmul(input=tensor1, other=tensor2) # tensor([18, 0, -53, 48])  tensor1 = torch.tensor([2, -5]) tensor2 = torch.tensor([[[3, 6, -1, 9], [-8, 0, 7, -2]], [[-7, -3, -4, 5], [-9, 4, -6, 0]]]) torch.matmul(input=tensor1, other=tensor2) # tensor([[46, 12, -37, 28], # [31, -26, 22, 10]])  tensor1 = torch.tensor([[2, -5], [4, 3]]) tensor2 = torch.tensor([[[3, 6, -1, 9], [-8, 0, 7, -2]], [[-7, -3, -4, 5], [-9, 4, -6, 0]]]) torch.matmul(input=tensor1, other=tensor2) # tensor([[[46, 12, -37, 28], # [-12, 24, 17, 30]], # [[31, -26, 22, 10], # [-55, 0, -34, 20]]])  tensor1 = torch.tensor([[2., -5.], [4., 3.]]) tensor2 = torch.tensor([[[3., 6., -1., 9.], [-8., 0., 7., -2.]], [[-7., -3., -4., 5.], [-9., 4., -6., 0.]]]) torch.matmul(input=tensor1, other=tensor2) # tensor([[[46., 12., -37., 28.], # [-12., 24., 17., 30.]], # [[31., -26., 22., 10.], # [-55., 0., -34., 20.]]])  tensor1 = torch.tensor([[2.+0.j, -5.+0.j], [4.+0.j, 3.+0.j]]) tensor2 = torch.tensor([[[3.+0.j, 6.+0.j, -1.+0.j, 9.+0.j], [-8.+0.j, 0.+0.j, 7.+0.j, -2.+0.j]], [[-7.+0.j, -3.+0.j, -4.+0.j, 5.+0.j], [-9.+0.j, 4.+0.j, -6.+0.j, 0.+0.j]]]) torch.matmul(input=tensor1, other=tensor2) # tensor([[[46.+0.j, 12.+0.j, -37.+0.j, 28.+0.j], # [-12.+0.j, 24.+0.j, 17.+0.j, 30.+0.j]], # [[31.+0.j, -26.+0.j, 22.+0.j, 10.+0.j], # [-55.+0.j, 0.+0.j, -34.+0.j, 20.+0.j]]])  tensor1 = torch.tensor([]) tensor2 = torch.tensor([]) torch.matmul(input=tensor1, other=tensor2) # tensor(0.) 
Enter fullscreen mode Exit fullscreen mode

dot() can do dot multiplication with two of the 1D tensors of zero or more elements, getting the 0D tensor of one element:

*Memos:

  • dot() can be used with torch or a tensor.
  • The 1st argument(input) with torch or using a tensor(Required-Type:tensor of int, float or complex). *It must be a 1D tesnor.
  • The 2nd argument with torch or the 1st argument with a tensor is tensor(Required-Type:tensor of int, float or complex). *It must be a 1D tesnor.
  • There is out argument with torch(Optional-Default:None-Type:tensor): *Memos:
    • out= must be used.
    • My post explains out argument.
import torch tensor1 = torch.tensor([2, -5, 4]) tensor2 = torch.tensor([3, 6, -1]) torch.dot(input=tensor1, tensor=tensor2) tensor1.dot(tensor=tensor2) # tensor(-28)  tensor1 = torch.tensor([2., -5., 4.]) tensor2 = torch.tensor([3., 6., -1.]) torch.dot(input=tensor1, tensor=tensor2) # tensor(-28.)  tensor1 = torch.tensor([2.+0.j, -5.+0.j, 4.+0.j]) tensor2 = torch.tensor([3.+0.j, 6.+0.j, -1.+0.j]) torch.dot(input=tensor1, tensor=tensor2) # tensor(-28.+0.j)  tensor1 = torch.tensor([]) tensor2 = torch.tensor([]) torch.dot(input=tensor1, tensor=tensor2) # tensor(0.) 
Enter fullscreen mode Exit fullscreen mode

Top comments (0)