fix csrtensor/cootensor docs

docs/api/api_python/mindspore/mindspore.COOTensor.rst

@@ -5,6 +5,16 @@ mindspore.COOTensor
 
     用来表示某一张量在给定索引上非零元素的集合，其中索引(indices)指示了每一个非零元素的位置。
 
+    对一个稠密Tensor `dense` 来说，它对应的COOTensor(indices, values, dense_shape)，满足 `dense[indices[i]] = values[i]` 。
+
+    如果 `indices` 是[[0, 1], [1, 2]]， `values` 是[1, 2]， `dense_shape` 是(3, 4)，那么它对应的稠密Tensor如下：
+
+    .. code-block::
+
+        [[0, 1, 0, 0],
+         [0, 0, 2, 0],
+         [0, 0, 0, 0]]
+
     .. note::
         这是一个实验特性，在未来可能会发生API的变化。
 

mindspore/python/mindspore/common/tensor.py

@@ -2585,11 +2585,27 @@ class COOTensor(COOTensor_):
         return COOTensor(self.indices, data, self.shape)
 
     def to_tuple(self):
-        """Return indices, values and shape as a tuple."""
+        """
+        Return indices, values and shape as a tuple.
+
+        Returns:
+            Tuple.
+
+        Supported Platforms:
+            ``Ascend`` ``GPU`` ``CPU``
+        """
         return self.indices, self.values, self.shape
 
     def abs(self):
-        """Return absolute value element-wisely."""
+        """
+        Return absolute value element-wisely.
+
+        Returns:
+            COOTensor.
+
+        Supported Platforms:
+            ``Ascend`` ``GPU`` ``CPU``
+        """
         data = self.values.abs()
         return COOTensor(self.indices, data, self.shape)
 
@@ -2709,7 +2725,15 @@ class CSRTensor(CSRTensor_):
         return len(self.shape)
 
     def to_tuple(self):
-        """Return indptr, indices, values and shape as a tuple."""
+        """
+        Return indptr, indices, values and shape as a tuple.
+
+        Returns:
+            Tuple.
+
+        Supported Platforms:
+            ``Ascend`` ``GPU`` ``CPU``
+        """
         return self.indptr, self.indices, self.values, self.shape
 
     def to_coo(self):