QuantizedGraphModule

class torch.ao.quantization.fx.graph_module.QuantizedGraphModule(*args, **kwargs)

This class is created to make sure PackedParams (e.g. LinearPackedParams, Conv2dPackedParams) to appear in state_dict so that we can serialize and deserialize quantized graph module with torch.save(m.state_dict()) and m.load_state_dict(state_dict)