FixedQParamsFakeQuantize

class torch.quantization.fake_quantize.FixedQParamsFakeQuantize(observer)

Simulate quantize and dequantize with fixed quantization parameters in training time. Only per tensor quantization is supported.