strided_slice

paddle.fluid.layers.nn. strided_slice ( input, axes, starts, ends, strides ) [source]
Alias_main

paddle.strided_slice :alias: paddle.strided_slice,paddle.tensor.strided_slice,paddle.tensor.manipulation.strided_slice :old_api: paddle.fluid.layers.strided_slice

This operator produces a slice of input along multiple axes. Similar to numpy: https://docs.scipy.org/doc/numpy/reference/arrays.indexing.html Slice uses axes, starts and ends attributes to specify the start and end dimension for each axis in the list of axes and Slice uses this information to slice the input data tensor. If a negative value is passed to starts or ends such as \(-i\), it represents the reverse position of the axis \(i-1\) th(here 0 is the initial position). The strides represents steps of slicing and if the strides is negative, slice operation is in the opposite direction. If the value passed to starts or ends is greater than n (the number of elements in this dimension), it represents n. For slicing to the end of a dimension with unknown size, it is recommended to pass in INT_MAX. The size of axes must be equal to starts , ends and strides. Following examples will explain how strided_slice works:

Case1:
    Given:
        data = [ [1, 2, 3, 4], [5, 6, 7, 8], ]
        axes = [0, 1]
        starts = [1, 0]
        ends = [2, 3]
        strides = [1, 1]
    Then:
        result = [ [5, 6, 7], ]

Case2:
    Given:
        data = [ [1, 2, 3, 4], [5, 6, 7, 8], ]
        axes = [0, 1]
        starts = [0, 1]
        ends = [2, 0]
        strides = [1, -1]
    Then:
        result = [ [8, 7, 6], ]

Case3:
    Given:
        data = [ [1, 2, 3, 4], [5, 6, 7, 8], ]
        axes = [0, 1]
        starts = [0, 1]
        ends = [-1, 1000]
        strides = [1, 3]
    Then:
        result = [ [2], ]

System Message: WARNING/2 (/usr/local/lib/python3.8/site-packages/paddle/fluid/layers/nn.py:docstring of paddle.fluid.layers.nn.strided_slice, line 50)

Explicit markup ends without a blank line; unexpected unindent.

Parameters

  • input (Variable) – An N-D Tensor or LoDTensor . The data type is bool, float32, float64, int32 or int64.

  • axes (list|tuple) – The data type is int32 . Axes that starts and ends apply to. It’s optional. If it is not provides, it will be treated as \([0,1,...,len(starts)-1]\).

  • starts (list|tuple|Variable) – The data type is int32 . If starts is a list or tuple, the elements of it should be integers or Tensors with shape [1]. If starts is an Variable, it should be an 1-D Tensor. It represents starting indices of corresponding axis in axes.

  • ends (list|tuple|Variable) – The data type is int32 . If ends is a list or tuple, the elements of it should be integers or Tensors with shape [1]. If ends is an Variable, it should be an 1-D Tensor . It represents ending indices of corresponding axis in axes.

  • strides (list|tuple|Variable) – The data type is int32 . If strides is a list or tuple, the elements of it should be integers or Tensors with shape [1]. If strides is an Variable, it should be an 1-D Tensor . It represents slice step of corresponding axis in axes.

Returns

A Tensor or LoDTensor with the same dimension as input. The data type is same as input.

Return type

Variable

Raises

  • TypeError – The type of starts must be list, tuple or Variable.

  • TypeError – The type of ends must be list, tuple or Variable.

  • TypeError – The type of strides must be list, tuple or Variable.

Examples

import paddle.fluid as fluid
import paddle

paddle.enable_static()
input = fluid.data(
    name="input", shape=[3, 4, 5, 6], dtype='float32')

# example 1:
# attr starts is a list which doesn't contain tensor Variable.
axes = [0, 1, 2]
starts = [-3, 0, 2]
ends = [3, 2, 4]
strides_1 = [1, 1, 1]
strides_2 = [1, 1, 2]
sliced_1 = fluid.layers.strided_slice(input, axes=axes, starts=starts, ends=ends, strides=strides_1)
# sliced_1 is input[:, 0:3:1, 0:2:1, 2:4:1].


# example 2:
# attr starts is a list which contain tensor Variable.
minus_3 = fluid.layers.fill_constant([1], "int32", -3)
sliced_2 = fluid.layers.strided_slice(input, axes=axes, starts=[minus_3, 0, 2], ends=ends, strides=strides_2)
# sliced_2 is input[:, 0:3:1, 0:2:1, 2:4:2].