How to do partial quantization by FX

torch.compile FX (Functional Transformations)
1

For eager mode, I can insert QuantStub and DeQuantStub nodes in the part that needs to be quantized, then for FX tool, how to realize partial quantization of the model?

1 Like