torch.bmm¶
- torch.bmm(input, mat2, *, out=None) Tensor¶
- Performs a batch matrix-matrix product of matrices stored in - inputand- mat2.- inputand- mat2must be 3-D tensors each containing the same number of matrices.- If - inputis a tensor,- mat2is a tensor,- outwill be a tensor.- This operator supports TensorFloat32. - On certain ROCm devices, when using float16 inputs this module will use different precision for backward. - Note - This function does not broadcast. For broadcasting matrix products, see - torch.matmul().- Parameters
- Keyword Arguments
- out (Tensor, optional) – the output tensor. 
 - Example: - >>> input = torch.randn(10, 3, 4) >>> mat2 = torch.randn(10, 4, 5) >>> res = torch.bmm(input, mat2) >>> res.size() torch.Size([10, 3, 5])