mindspore 2.1.0__cp38-cp38-manylinux1_x86_64.whl → 2.2.11__cp38-cp38-manylinux1_x86_64.whl

mindspore/nn/layer/rnns.py

@@ -372,7 +372,7 @@ class _RNNBase(Cell):
     '''Basic class for RNN operators'''
 
     def __init__(self, mode, input_size, hidden_size, num_layers=1, has_bias=True,
-                 batch_first=False, dropout=0., bidirectional=False):
+                 batch_first=False, dropout=0., bidirectional=False, dtype=mstype.float32):
         super().__init__()
         validator.check_positive_int(hidden_size, "hidden_size", self.cls_name)
         validator.check_positive_int(input_size, "input_size", self.cls_name)
@@ -435,17 +435,17 @@ class _RNNBase(Cell):
                 suffix = '_reverse' if direction == 1 else ''
 
                 self.w_ih_list.append(Parameter(
-                    Tensor(np.random.uniform(-stdv, stdv, (gate_size, layer_input_size)).astype(np.float32)),
-                    name='weight_ih_l{}{}'.format(layer, suffix)))
+                    Tensor(np.random.uniform(-stdv, stdv, (gate_size, layer_input_size)).astype(np.float32),
+                           dtype=dtype), name='weight_ih_l{}{}'.format(layer, suffix)))
                 self.w_hh_list.append(Parameter(
-                    Tensor(np.random.uniform(-stdv, stdv, (gate_size, hidden_size)).astype(np.float32)),
-                    name='weight_hh_l{}{}'.format(layer, suffix)))
+                    Tensor(np.random.uniform(-stdv, stdv, (gate_size, hidden_size)).astype(np.float32),
+                           dtype=dtype), name='weight_hh_l{}{}'.format(layer, suffix)))
                 if has_bias:
                     self.b_ih_list.append(Parameter(
-                        Tensor(np.random.uniform(-stdv, stdv, (gate_size)).astype(np.float32)),
+                        Tensor(np.random.uniform(-stdv, stdv, (gate_size)).astype(np.float32), dtype=dtype),
                         name='bias_ih_l{}{}'.format(layer, suffix)))
                     self.b_hh_list.append(Parameter(
-                        Tensor(np.random.uniform(-stdv, stdv, (gate_size)).astype(np.float32)),
+                        Tensor(np.random.uniform(-stdv, stdv, (gate_size)).astype(np.float32), dtype=dtype),
                         name='bias_hh_l{}{}'.format(layer, suffix)))
         self.w_ih_list = ParameterTuple(self.w_ih_list)
         self.w_hh_list = ParameterTuple(self.w_hh_list)
@@ -578,9 +578,7 @@ class _RNNBase(Cell):
 
 class RNN(_RNNBase):
     r"""
-    Stacked Elman RNN layers.
-
-    Apply RNN layer with :math:`\tanh` or :math:`\text{ReLU}` non-linearity to the input.
+    Stacked Elman RNN layers, applying RNN layer with :math:`\tanh` or :math:`\text{ReLU}` non-linearity to the input.
 
     For each element in the input sequence, each layer computes the following function:
 
@@ -597,19 +595,19 @@ class RNN(_RNNBase):
         hidden_size (int):  Number of features of hidden layer.
         num_layers (int): Number of layers of stacked RNN. Default: ``1`` .
         nonlinearity (str): The non-linearity to use. Can be either ``'tanh'`` or ``'relu'``. Default: ``'tanh'``
-        has_bias (bool): Whether the cell has bias `b_ih` and `b_hh`. Default: ``True`` .
+        has_bias (bool): Whether the cell has bias :math:`b_{ih}` and :math:`b_{hh}`. Default: ``True`` .
         batch_first (bool): Specifies whether the first dimension of input `x` is batch_size. Default: ``False`` .
         dropout (float): If not 0.0, append `Dropout` layer on the outputs of each
             RNN layer except the last layer. Default ``0.0`` . The range of dropout is [0.0, 1.0).
         bidirectional (bool): Specifies whether it is a bidirectional RNN,
             num_directions=2 if bidirectional=True otherwise 1. Default: ``False`` .
+        dtype (:class:`mindspore.dtype`): Dtype of Parameters. Default: ``mstype.float32`` .
 
     Inputs:
         - **x** (Tensor) - Tensor of data type mindspore.float32 or mindspore.float16 and
           shape :math:`(seq\_len, batch\_size, input\_size)` or :math:`(batch\_size, seq\_len, input\_size)` .
         - **hx** (Tensor) - Tensor of data type mindspore.float32 or mindspore.float16 and
           shape :math:`(num\_directions * num\_layers, batch\_size, hidden\_size)` .
-          The data type of `hx` must be the same as `x`.
         - **seq_length** (Tensor) - The length of each sequence in an input batch.
           Tensor of shape :math:`(batch\_size)` . Default: ``None`` .
           This input indicates the real sequence length before padding to avoid padded elements
@@ -704,14 +702,13 @@ class GRU(_RNNBase):
             GRU layer except the last layer. Default ``0.0`` . The range of dropout is [0.0, 1.0).
         bidirectional (bool): Specifies whether it is a bidirectional GRU,
             num_directions=2 if bidirectional=True otherwise 1. Default: ``False`` .
+        dtype (:class:`mindspore.dtype`): Dtype of Parameters. Default: ``mstype.float32`` .
 
     Inputs:
         - **x** (Tensor) - Tensor of data type mindspore.float32 or mindspore.float16 and
           shape (seq\_len, batch\_size, `input\_size`) or :math:`(batch\_size, seq\_len, input\_size)`.
         - **hx** (Tensor) - Tensor of data type mindspore.float32 or mindspore.float16 and
           shape :math:`(num\_directions * num\_layers, batch\_size, hidden\_size)`.
-          The data type of `hx` must be the same as
-          `x`.
         - **seq_length** (Tensor) - The length of each sequence in an input batch.
           Tensor of shape :math:`(\text{batch_size})`. Default: ``None`` .
           This input indicates the real sequence length before padding to avoid padded elements
@@ -799,19 +796,19 @@ class LSTM(_RNNBase):
         input_size (int): Number of features of input.
         hidden_size (int):  Number of features of hidden layer.
         num_layers (int): Number of layers of stacked LSTM . Default: ``1`` .
-        has_bias (bool): Whether the cell has bias `b_ih` and `b_hh`. Default: ``True`` .
+        has_bias (bool): Whether the cell has bias :math:`b_{ih}` and :math:`b_{hh}`. Default: ``True`` .
         batch_first (bool): Specifies whether the first dimension of input `x` is batch_size. Default: ``False`` .
         dropout (float, int): If not 0, append `Dropout` layer on the outputs of each
             LSTM layer except the last layer. Default ``0`` . The range of dropout is [0.0, 1.0).
         bidirectional (bool): Specifies whether it is a bidirectional LSTM,
             num_directions=2 if bidirectional=True otherwise 1. Default: ``False`` .
+        dtype (:class:`mindspore.dtype`): Dtype of Parameters. Default: ``mstype.float32`` .
 
     Inputs:
         - **x** (Tensor) - Tensor of data type mindspore.float32 or mindspore.float16 and
           shape :math:`(seq\_len, batch\_size, input\_size)` or :math:`(batch\_size, seq\_len, input\_size)` .
         - **hx** (tuple) - A tuple of two Tensors (h_0, c_0) both of data type mindspore.float32
           or mindspore.float16 and shape :math:`(num\_directions * num\_layers, batch\_size, hidden\_size)` .
-          The data type of `hx` must be the same as `x`.
         - **seq_length** (Tensor) - The length of each sequence in an input batch.
           Tensor of shape :math:`(batch\_size)`. Default: ``None`` .
           This input indicates the real sequence length before padding to avoid padded elements

mindspore/nn/layer/thor_layer.py

@@ -602,7 +602,7 @@ class EmbeddingThor(Cell):
         self.expand = P.ExpandDims()
         self.reshape_flat = P.Reshape()
         self.shp_flat = (-1,)
-        self.gather = P.GatherV2()
+        self.gather = P.Gather()
         self.one_hot = P.OneHot()
         self.on_value = Tensor(1.0, self.dtype)
         self.off_value = Tensor(0.0, self.dtype)