Rate this Page

hardswish#

class torch.ao.nn.quantized.functional.hardswish(input, scale, zero_point)[source]#

This is the quantized version of hardswish().

Parameters
  • input (Tensor) – quantized input

  • scale (float) – quantization scale of the output tensor

  • zero_point (int) – quantization zero point of the output tensor

Return type

Tensor