MCPcopy Create free account
hub / github.com/DeepRec-AI/DeepRec / avg_pool

Function avg_pool

tensorflow/python/ops/nn_ops.py:3642–3682  ·  view source on GitHub ↗

Performs the average pooling on the input. Each entry in `output` is the mean of the corresponding size `ksize` window in `value`. Args: value: A 4-D `Tensor` of shape `[batch, height, width, channels]` and type `float32`, `float64`, `qint8`, `quint8`, or `qint32`. ksize: An in

(value, ksize, strides, padding, data_format="NHWC",
             name=None, input=None)

Source from the content-addressed store, hash-verified

3640
3641@tf_export(v1=["nn.avg_pool", "nn.avg_pool2d"])
3642def avg_pool(value, ksize, strides, padding, data_format="NHWC",
3643 name=None, input=None): # pylint: disable=redefined-builtin
3644 """Performs the average pooling on the input.
3645
3646 Each entry in `output` is the mean of the corresponding size `ksize`
3647 window in `value`.
3648
3649 Args:
3650 value: A 4-D `Tensor` of shape `[batch, height, width, channels]` and type
3651 `float32`, `float64`, `qint8`, `quint8`, or `qint32`.
3652 ksize: An int or list of `ints` that has length `1`, `2` or `4`. The size of
3653 the window for each dimension of the input tensor.
3654 strides: An int or list of `ints` that has length `1`, `2` or `4`. The
3655 stride of the sliding window for each dimension of the input tensor.
3656 padding: A string, either `'VALID'` or `'SAME'`. The padding algorithm.
3657 See the "returns" section of `tf.nn.convolution` for details.
3658 data_format: A string. 'NHWC' and 'NCHW' are supported.
3659 name: Optional name for the operation.
3660 input: Alias for value.
3661
3662 Returns:
3663 A `Tensor` with the same type as `value`. The average pooled output tensor.
3664 """
3665 with ops.name_scope(name, "AvgPool", [value]) as name:
3666 value = deprecation.deprecated_argument_lookup(
3667 "input", input, "value", value)
3668
3669 if data_format is None:
3670 data_format = "NHWC"
3671 channel_index = 1 if data_format.startswith("NC") else 3
3672
3673 ksize = _get_sequence(ksize, 2, channel_index, "ksize")
3674 strides = _get_sequence(strides, 2, channel_index, "strides")
3675
3676 return gen_nn_ops.avg_pool(
3677 value,
3678 ksize=ksize,
3679 strides=strides,
3680 padding=padding,
3681 data_format=data_format,
3682 name=name)
3683
3684
3685@tf_export("nn.avg_pool2d", v1=[])

Callers

nothing calls this directly

Calls 2

_get_sequenceFunction · 0.70
name_scopeMethod · 0.45

Tested by

no test coverage detected