mindspore.ops.Concat

class mindspore.ops.Concat(axis=0)

Connect tensor in the specified axis.

Connect input tensors along with the given axis.

The input data is a tuple of tensors. These tensors have the same rank R. Set the given axis as m, and \(0 \le m < R\). Set the number of input tensors as N. For the \(i\)-th tensor \(t_i\), it has the shape of \((x_1, x_2, ..., x_{mi}, ..., x_R)\). \(x_{mi}\) is the \(m\)-th dimension of the \(i\)-th tensor. Then, the shape of the output tensor is

\[(x_1, x_2, ..., \sum_{i=1}^Nx_{mi}, ..., x_R)\]

Warning

The value range of “axis” is [-dims, dims - 1]. “dims” is the dimension length of “input_x”.

Parameters

axis (int) – The specified axis. Default: 0.

Inputs:

• input_x (tuple, list) - A tuple or a list of input tensors. Suppose there are two tensors in this tuple or list, namely x1 and x2. To perform Concat in the axis 0 direction, except for the 0th axis, all other axes should be equal, that is, \(x1.shape[1] == x2.shape[1], x1.shape[2] == x2.shape[2], ..., x1.shape[R] == x2.shape[R]\), where the \(R\) indicates the last axis.

Outputs:

• Tensor, the shape is \((x_1, x_2, ..., \sum_{i=1}^Nx_{mi}, ..., x_R)\). The data type is the same with input_x.

Raises

TypeError – If axis is not an int.

Supported Platforms:

Ascend GPU CPU

Examples

>>> input_x1 = Tensor(np.array([[0, 1], [2, 1]]).astype(np.float32))
>>> input_x2 = Tensor(np.array([[0, 1], [2, 1]]).astype(np.float32))
>>> op = ops.Concat()
>>> output = op((input_x1, input_x2))
>>> print(output)
[[0. 1.]
 [2. 1.]
 [0. 1.]
 [2. 1.]]
>>> op = ops.Concat(1)
>>> output = op((input_x1, input_x2))
>>> print(output)
[[0. 1. 0. 1.]
 [2. 1. 2. 1.]]