int_scaled_matmul

torchao.quantization.int_scaled_matmul(a: Tensor, b: Tensor, scales1: Tensor) → Tensor

Performs scaled integer matrix multiplication.

Parameters

• a (torch.Tensor) – The first matrix to multiply.

• b (torch.Tensor) – The second matrix to multiply.

• scales1 (torch.Tensor) – The scaling factors for the rows of the result.

Returns

The result of the scaled matrix multiplication.

Return type

torch.Tensor

Raises

AssertionError – If the dimensions of the input tensors do not match the expected shapes.