!2917 bug fix in quantization aware training auto create graph

Merge pull request !2917 from chenzhongming/master

mindspore/nn/layer/quant.py

@@ -855,7 +855,7 @@ class ActQuant(_QuantActivation):
                                                   symmetric=symmetric,
                                                   narrow_range=narrow_range,
                                                   quant_delay=quant_delay)
-        self.act = activation
+        self.act = activation()
 
     def construct(self, x):
         x = self.act(x)
@@ -921,7 +921,7 @@ class HSwishQuant(_QuantActivation):
                                                         narrow_range=narrow_range,
                                                         quant_delay=quant_delay)
         if isinstance(activation, nn.HSwish):
-            self.act = activation
+            self.act = activation()
         else:
             raise ValueError("Activation should be `nn.HSwish`")
 
@@ -990,7 +990,7 @@ class HSigmoidQuant(_QuantActivation):
                                                         narrow_range=narrow_range,
                                                         quant_delay=quant_delay)
         if isinstance(activation, nn.HSwish):
-            self.act = activation
+            self.act = activation()
         else:
             raise ValueError("Activation should be `nn.HSigmoid`")
 

mindspore/train/quant/quant.py

@@ -114,7 +114,6 @@ class ConvertToQuantNetwork:
     def run(self):
         self.network.update_cell_prefix()
         network = self._convert_subcells2quant(self.network)
-        network = _AddFakeQuantInput(network)
         self.network.update_cell_type("quant")
         return network
 
@@ -275,16 +274,20 @@ class ExportToQuantInferNetwork:
     Args:
         network (Cell): MindSpore network API `convert_quant_network`.
         inputs (Tensor): Input tensors of the `quantization aware training network`.
+        mean (int): Input data mean. Default: 127.5.
+        std_dev (int, float): Input data variance. Default: 127.5.
 
     Returns:
         Cell, GEIR backend Infer network.
     """
     __quant_op_name__ = ["TensorAdd", "Sub", "Mul", "RealDiv"]
 
-    def __init__(self,
-                 network,
-                 *inputs):
+    def __init__(self, network, mean, std_dev, *inputs):
         network = validator.check_isinstance('network', network, (nn.Cell,))
+        # quantize for inputs: q = f / scale + zero_point
+        # dequantize for outputs: f = (q - zero_point) * scale
+        self.input_scale = round(mean)
+        self.input_zero_point = 1 / std_dev
         self.data_type = mstype.int8
         self.network = copy.deepcopy(network)
         self.all_parameters = {p.name: p for p in self.network.get_parameters()}
@@ -395,7 +398,7 @@ class ExportToQuantInferNetwork:
         return network
 
 
-def export(network, *inputs, file_name, file_format='GEIR'):
+def export(network, *inputs, file_name, mean=127.5, std_dev=127.5, file_format='GEIR'):
     """
     Exports MindSpore quantization predict model to deploy with GEIR.
 
@@ -403,12 +406,17 @@ def export(network, *inputs, file_name, file_format='GEIR'):
         network (Cell): MindSpore network produced by `convert_quant_network`.
         inputs (Tensor): Inputs of the `quantization aware training network`.
         file_name (str): File name of model to export.
+        mean (int): Input data mean. Default: 127.5.
+        std_dev (int, float): Input data variance. Default: 127.5.
         file_format (str): MindSpore currently supports 'GEIR' format for exported quantization aware model.
             - GEIR: Graph Engine Intermediate Representation. An Intermediate representation format of Ascend model.
     """
     supported_device = ["Ascend"]
     supported_formats = ['GEIR']
 
+    mean = validator.check_type("mean", mean, (int, float))
+    std_dev = validator.check_type("std_dev", std_dev, (int, float))
+
     if context.get_context('device_target') not in supported_device:
         raise KeyError("Unsupported {} device target.".format(context.get_context('device_target')))
 
@@ -418,7 +426,7 @@ def export(network, *inputs, file_name, file_format='GEIR'):
     network.set_train(False)
 
     if file_format == 'GEIR':
-        exporter = ExportToQuantInferNetwork(network, *inputs)
+        exporter = ExportToQuantInferNetwork(network, mean, std_dev, *inputs)
         deploy_net = exporter.run()
         serialization.export(deploy_net, *inputs, file_name=file_name, file_format=file_format)