Method quantize_large_weights

tools/python/transform/transformer.py:2143–2149

(self)

                    self._quantized_tensor.update([tensor.name])

    def quantize_large_weights(self):
        print("Quantize large weights")
        net = self._model
        for tensor in net.tensors:
            self.quantize_large_tensor(tensor)

        return False

    def add_quantize_info(self, op, minval, maxval):
        quantize_schema = self._option.quantize_schema