* add doc for some Parameter property.* add export in quant __all__* fix bug for isinstance & add ut for it

mindspore/common/dtype.py

@@ -34,6 +34,8 @@ __dtype__ = [
     "float64", "double",
     "bool_", "float_",
     "list_", "tuple_",
+    "int_", "uint",
+    "number", "tensor",
     "string", "type_none",
     "tensor_type",
     "Type"

mindspore/common/parameter.py

@@ -180,7 +180,12 @@ class Parameter(MetaTensor):
 
     @property
     def inited_param(self):
-        """Get the new parameter after call the init_data."""
+        """
+        Get the new parameter after call the init_data.
+
+        Default is a None, If `self` is a Parameter with out data, after call the
+        `init_data` the initialized Parameter with data will be recorded here.
+        """
         return self._inited_param
 
 
@@ -232,7 +237,14 @@ class Parameter(MetaTensor):
 
     @property
     def is_init(self):
-        """Get the initialization status of the parameter."""
+        """
+        Get the initialization status of the parameter.
+
+        In GE backend, the Parameter need a "init graph" to sync the data from host to device.
+        This flag indicates whether the data as been sync to the device.
+
+        This flag only work in GE, and it will be set to False in other backend.
+        """
         return self._is_init
 
     @is_init.setter
@@ -250,7 +262,8 @@ class Parameter(MetaTensor):
         Clone the parameter.
 
         Args:
-            prefix (str): Namespace of parameter.
+            prefix (str): Namespace of parameter. The cloned Parameter name is
+                combined of prefix and current name: `f"{perfix}.{self.name}"`.
             init (Union[Tensor, str, Initializer, numbers.Number]): Initialize the shape of the parameter.
                 Default: 'same'.
 

mindspore/core/ir/dtype/number.h

@@ -92,7 +92,12 @@ class Int : public Number {
   ~Int() override = default;
   MS_DECLARE_PARENT(Int, Number)
   TypeId generic_type_id() const override { return kNumberTypeInt; }
-  TypePtr DeepCopy() const override { return std::make_shared<Int>(nbits()); }
+  TypePtr DeepCopy() const override {
+    if (nbits() == 0) {
+      return std::make_shared<Int>();
+    }
+    return std::make_shared<Int>(nbits());
+  }
   std::string ToString() const override { return GetTypeName("Int"); }
   std::string ToReprString() const override { return nbits() == 0 ? "int_" : GetTypeName("int"); }
   std::string DumpText() const override {
@@ -110,7 +115,12 @@ class UInt : public Number {
   ~UInt() override {}
   MS_DECLARE_PARENT(UInt, Number)
 
-  TypePtr DeepCopy() const override { return std::make_shared<UInt>(nbits()); }
+  TypePtr DeepCopy() const override {
+    if (nbits() == 0) {
+      return std::make_shared<UInt>();
+    }
+    return std::make_shared<UInt>(nbits());
+  }
   std::string ToString() const override { return GetTypeName("UInt"); }
   std::string ToReprString() const override { return GetTypeName("uint"); }
   std::string DumpText() const override {
@@ -127,7 +137,12 @@ class Float : public Number {
   MS_DECLARE_PARENT(Float, Number)
 
   TypeId generic_type_id() const override { return kNumberTypeFloat; }
-  TypePtr DeepCopy() const override { return std::make_shared<Float>(nbits()); }
+  TypePtr DeepCopy() const override {
+    if (nbits() == 0) {
+      return std::make_shared<Float>();
+    }
+    return std::make_shared<Float>(nbits());
+  }
   std::string ToString() const override { return GetTypeName("Float"); }
   std::string ToReprString() const override { return nbits() == 0 ? "float_" : GetTypeName("float"); }
   std::string DumpText() const override {

mindspore/ops/operations/array_ops.py

@@ -349,7 +349,12 @@ class IsInstance(PrimitiveWithInfer):
         validator.check_const_input("inst", inst['value'], self.name)
         validator.check_value_type("type_", type_v, [mstype.Type], self.name)
 
-        value = mstype.issubclass_(sub_type_t, type_v)
+        if type_v == mstype.list_:
+            value = isinstance(sub_type_t, list)
+        elif type_v == mstype.tuple_:
+            value = isinstance(sub_type_t, tuple)
+        else:
+            value = mstype.issubclass_(sub_type_t, type_v)
 
         out = {'shape': (),
                'dtype': mstype.type_type,

mindspore/train/model.py

@@ -70,11 +70,14 @@ class Model:
                              value would be passed to the `Loss` metric, the predicted value and label would be passed
                              to other metric. Default: None.
         amp_level (str): Option for argument `level` in `mindspore.amp.build_train_network`, level for mixed
-            precision training. Supports [O0, O2, O3]. Default: "O0".
+            precision training. Supports ["O0", "O2", "O3", "auto"]. Default: "O0".
 
             - O0: Do not change.
             - O2: Cast network to float16, keep batchnorm run in float32, using dynamic loss scale.
             - O3: Cast network to float16, with additional property 'keep_batchnorm_fp32=False'.
+            - auto: Set to level to recommended level in different devices. Set level to O2 on GPU, Set
+              level to O3 Ascend. The recommended level is choose by the export experience, cannot
+              always generalize. User should specify the level for special network.
 
             O2 is recommended on GPU, O3 is recommended on Ascend.
 

mindspore/train/quant/__init__.py

@@ -21,6 +21,6 @@ operations. Note that the entire computation is carried out in floating point. A
 aware training, MindSpore provides conversion functions to convert the trained model into lower precision.
 """
 
-from .quant import convert_quant_network
+from .quant import convert_quant_network, export
 
-__all__ = ["convert_quant_network"]
+__all__ = ["convert_quant_network", "export"]

tests/ut/python/pynative_mode/ops/test_isinstance.py

@@ -0,0 +1,24 @@
+# Copyright 2020 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+""" test_multitype """
+import mindspore as ms
+from mindspore.ops import operations as P
+
+
+def test_isinstance():
+    assert P.IsInstance()([1, 2, 3], ms.list_) is True
+    assert P.IsInstance()((1, 2, 3), ms.tuple_) is True
+    assert P.IsInstance()(1.0, ms.float_) is True
+    assert P.IsInstance()(1, ms.int_) is True