stack¶

paddle. stack ( x, axis=0 ) [源代码] ¶

该OP沿 axis 轴对输入 x 进行堆叠操作。

例1:

输入:
    x[0].shape = [1, 2]
    x[0].data = [ [1.0 , 2.0 ] ]
    x[1].shape = [1, 2]
    x[1].data = [ [3.0 , 4.0 ] ]
    x[2].shape = [1, 2]
    x[2].data = [ [5.0 , 6.0 ] ]

参数:
    axis = 0 #沿着第0维对输入x进行堆叠操作。

输出:
    Out.shape = [3, 1, 2]
    Out.data = [ [ [1.0, 2.0] ],
                [ [3.0, 4.0] ],
                [ [5.0, 6.0] ] ]

例2:

输入:
    x[0].shape = [1, 2]
    x[0].data = [ [1.0 , 2.0 ] ]
    x[1].shape = [1, 2]
    x[1].data = [ [3.0 , 4.0 ] ]
    x[2].shape = [1, 2]
    x[2].data = [ [5.0 , 6.0 ] ]

参数:
    axis = 1 or axis = -2 #沿着第1维对输入进行堆叠操作。

输出:
    Out.shape = [1, 3, 2]
    Out.data = [ [ [1.0, 2.0]
                  [3.0, 4.0]
                  [5.0, 6.0] ] ]

参数:

x (list(Variable)|tuple(Variable)) – 输入 x 可以是单个Tensor，或是多个Tensor组成的列表。如果 x 是一个列表，那么这些Tensor的维度必须相同。假设输入是N维Tensor \([d_0,d_1,...,d_{n−1}]\)，则输出变量的维度为N+1维 \([d_0,d_1,...d_{axis-1},len(x),d_{axis}...,d_{n−1}]\) 。支持的数据类型: float32，float64，int32，int64。
axis (int, 可选) – 指定对输入Tensor进行堆叠运算的轴，有效 axis 的范围是: \([-(R+1), R+1)\)，R是输入中第一个Tensor的rank。如果 axis < 0，则 \(axis=axis+rank(x[0])+1\) 。axis默认值为0。

返回: 堆叠运算后的Tensor，数据类型与输入Tensor相同。输出维度等于 \(rank(x[0])+1\) 维。

返回类型: Variable

代码示例：

import paddle.fluid as fluid
import paddle.fluid.layers as layers
# set batch size=None
x1 = fluid.data(name='x1', shape=[None, 1, 2], dtype='int32')
x2 = fluid.data(name='x2', shape=[None, 1, 2], dtype='int32')
# stack Tensor list
data = layers.stack([x1,x2]) # stack according to axis 0, data.shape=[2, None, 1, 2]

data = layers.stack([x1,x2], axis=1) # stack according to axis 1, data.shape=[None, 2, 1, 2]

PaddlePaddle

paddle.fluid / layers / stack

stack¶

PaddlePaddle

paddle

paddle.amp

paddle.callbacks

paddle.compat

paddle.distributed

paddle.distribution

paddle.fluid

paddle.io

paddle.jit

paddle.metric

paddle.nn

paddle.onnx

paddle.optimizer

paddle.regularizer

paddle.static

paddle.sysconfig

paddle.text

paddle.utils

paddle.vision

paddle.fluid / layers / stack

stack¶