add float16 check for gpu fakequant op

mindspore/ops/operations/_quant_ops.py

@@ -467,6 +467,9 @@ class FakeQuantPerLayer(PrimitiveWithInfer):
         return x_shape
 
     def infer_dtype(self, x_type, min_type, max_type):
+        if context.get_context('device_target') == "GPU":
+            valid_types = (mstype.float32,)
+        else:
             valid_types = (mstype.float16, mstype.float32)
         validator.check_tensor_type_same({"x": x_type}, valid_types, self.name)
         validator.check_tensor_type_same(
@@ -521,6 +524,9 @@ class FakeQuantPerLayerGrad(PrimitiveWithInfer):
         return dout_shape
 
     def infer_dtype(self, dout_type, x_type, min_type, max_type):
+        if context.get_context('device_target') == "GPU":
+            valid_types = (mstype.float32,)
+        else:
             valid_types = (mstype.float16, mstype.float32)
         validator.check_tensor_type_same(
             {"dout": dout_type}, valid_types, self.name)
@@ -616,6 +622,9 @@ class FakeQuantPerChannel(PrimitiveWithInfer):
         return x_shape
 
     def infer_dtype(self, x_type, min_type, max_type):
+        if context.get_context('device_target') == "GPU":
+            valid_types = (mstype.float32,)
+        else:
             valid_types = (mstype.float16, mstype.float32)
         validator.check_tensor_type_same({"x": x_type}, valid_types, self.name)
         validator.check_tensor_type_same(
@@ -670,6 +679,9 @@ class FakeQuantPerChannelGrad(PrimitiveWithInfer):
         return dout_shape
 
     def infer_dtype(self, dout_type, x_type, min_type, max_type):
+        if context.get_context('device_target') == "GPU":
+            valid_types = (mstype.float32,)
+        else:
             valid_types = (mstype.float16, mstype.float32)
         validator.check_tensor_type_same(
             {"dout": dout_type}, valid_types, self.name)