PyTorch中的矩阵乘法torch.mm()torch.bmm()torch.matmul()torch.mul()总结
torch.mm()torch.mm(input, mat2, out=None) → Tensor矩阵乘法,不进行 broadcasttorch.bmm()输入1 :(b×n×m) tensor, 输入2:(b×m×p) tensor, 输出:(b×n×p) tensor.batch 式的矩阵乘法,不broadcasttorch.matmul()torch.matmul(input, other, out=None) → Tensor矩阵乘法,有broadcast功能如果输入的tens