You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by jx...@apache.org on 2018/05/15 18:20:57 UTC
[incubator-mxnet] branch master updated: [MXNET-9704] An assertion
check for invalid layout (#10915)
This is an automated email from the ASF dual-hosted git repository.
jxie pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/master by this push:
new 0805e4d [MXNET-9704] An assertion check for invalid layout (#10915)
0805e4d is described below
commit 0805e4de3a1070763349fb245e00845af4ca8d49
Author: Yan Li <go...@gmail.com>
AuthorDate: Wed May 16 02:20:51 2018 +0800
[MXNET-9704] An assertion check for invalid layout (#10915)
* add assertion and update documentations for gluon conv and pooling layers
* Update conv_layers.py
fix pylint
* fix for pylint
---
python/mxnet/gluon/nn/conv_layers.py | 104 +++++++++++++++++++----------------
1 file changed, 57 insertions(+), 47 deletions(-)
diff --git a/python/mxnet/gluon/nn/conv_layers.py b/python/mxnet/gluon/nn/conv_layers.py
index 87a62bc..7b4a6be 100644
--- a/python/mxnet/gluon/nn/conv_layers.py
+++ b/python/mxnet/gluon/nn/conv_layers.py
@@ -196,7 +196,7 @@ class Conv1D(_Conv):
layers side by side, each seeing half the input channels, and producing
half the output channels, and both subsequently concatenated.
layout: str, default 'NCW'
- Dimension ordering of data and weight. Can be 'NCW', 'NWC', etc.
+ Dimension ordering of data and weight. Only supports 'NCW' layout for now.
'N', 'C', 'W' stands for batch, channel, and width (time) dimensions
respectively. Convolution is applied on the 'W' dimension.
in_channels : int, default 0
@@ -229,6 +229,7 @@ class Conv1D(_Conv):
groups=1, layout='NCW', activation=None, use_bias=True,
weight_initializer=None, bias_initializer='zeros',
in_channels=0, **kwargs):
+ assert layout == 'NCW', "Only supports 'NCW' layout for now"
if isinstance(kernel_size, numeric_types):
kernel_size = (kernel_size,)
assert len(kernel_size) == 1, "kernel_size must be a number or a list of 1 ints"
@@ -271,9 +272,9 @@ class Conv2D(_Conv):
layers side by side, each seeing half the input channels, and producing
half the output channels, and both subsequently concatenated.
layout : str, default 'NCHW'
- Dimension ordering of data and weight. Can be 'NCHW', 'NHWC', etc.
- 'N', 'C', 'H', 'W' stands for batch, channel, height, and width
- dimensions respectively. Convolution is applied on the 'H' and
+ Dimension ordering of data and weight. Only supports 'NCHW' and 'NHWC'
+ layout for now. 'N', 'C', 'H', 'W' stands for batch, channel, height,
+ and width dimensions respectively. Convolution is applied on the 'H' and
'W' dimensions.
in_channels : int, default 0
The number of input channels to this layer. If not specified,
@@ -293,12 +294,12 @@ class Conv2D(_Conv):
Inputs:
- **data**: 4D input tensor with shape
- `(batch_size, in_channels, height, width)` when `layout` is `NCW`.
+ `(batch_size, in_channels, height, width)` when `layout` is `NCHW`.
For other layouts shape is permuted accordingly.
Outputs:
- **out**: 4D output tensor with shape
- `(batch_size, channels, out_height, out_width)` when `layout` is `NCW`.
+ `(batch_size, channels, out_height, out_width)` when `layout` is `NCHW`.
out_height and out_width are calculated as::
out_height = floor((height+2*padding[0]-dilation[0]*(kernel_size[0]-1)-1)/stride[0])+1
@@ -308,6 +309,8 @@ class Conv2D(_Conv):
dilation=(1, 1), groups=1, layout='NCHW',
activation=None, use_bias=True, weight_initializer=None,
bias_initializer='zeros', in_channels=0, **kwargs):
+ assert layout == 'NCHW' or layout == 'NHWC', \
+ "Only supports 'NCHW' and 'NHWC' layout for now"
if isinstance(kernel_size, numeric_types):
kernel_size = (kernel_size,)*2
assert len(kernel_size) == 2, "kernel_size must be a number or a list of 2 ints"
@@ -350,9 +353,9 @@ class Conv3D(_Conv):
layers side by side, each seeing half the input channels, and producing
half the output channels, and both subsequently concatenated.
layout : str, default 'NCDHW'
- Dimension ordering of data and weight. Can be 'NCDHW', 'NDHWC', etc.
- 'N', 'C', 'H', 'W', 'D' stands for batch, channel, height, width and
- depth dimensions respectively. Convolution is applied on the 'D',
+ Dimension ordering of data and weight. Only supports 'NCDHW' and 'NDHWC'
+ layout for now. 'N', 'C', 'H', 'W', 'D' stands for batch, channel, height,
+ width and depth dimensions respectively. Convolution is applied on the 'D',
'H' and 'W' dimensions.
in_channels : int, default 0
The number of input channels to this layer. If not specified,
@@ -372,12 +375,12 @@ class Conv3D(_Conv):
Inputs:
- **data**: 5D input tensor with shape
- `(batch_size, in_channels, depth, height, width)` when `layout` is `NCW`.
+ `(batch_size, in_channels, depth, height, width)` when `layout` is `NCDHW`.
For other layouts shape is permuted accordingly.
Outputs:
- **out**: 5D output tensor with shape
- `(batch_size, channels, out_depth, out_height, out_width)` when `layout` is `NCW`.
+ `(batch_size, channels, out_depth, out_height, out_width)` when `layout` is `NCDHW`.
out_depth, out_height and out_width are calculated as::
out_depth = floor((depth+2*padding[0]-dilation[0]*(kernel_size[0]-1)-1)/stride[0])+1
@@ -388,6 +391,8 @@ class Conv3D(_Conv):
dilation=(1, 1, 1), groups=1, layout='NCDHW', activation=None,
use_bias=True, weight_initializer=None, bias_initializer='zeros',
in_channels=0, **kwargs):
+ assert layout == 'NCDHW' or layout == 'NDHWC', \
+ "Only supports 'NCDHW' and 'NDHWC' layout for now"
if isinstance(kernel_size, numeric_types):
kernel_size = (kernel_size,)*3
assert len(kernel_size) == 3, "kernel_size must be a number or a list of 3 ints"
@@ -431,7 +436,7 @@ class Conv1DTranspose(_Conv):
layers side by side, each seeing half the input channels, and producing
half the output channels, and both subsequently concatenated.
layout : str, default 'NCW'
- Dimension ordering of data and weight. Can be 'NCW', 'NWC', etc.
+ Dimension ordering of data and weight. Only supports 'NCW' layout for now.
'N', 'C', 'W' stands for batch, channel, and width (time) dimensions
respectively. Convolution is applied on the 'W' dimension.
in_channels : int, default 0
@@ -464,6 +469,7 @@ class Conv1DTranspose(_Conv):
dilation=1, groups=1, layout='NCW', activation=None, use_bias=True,
weight_initializer=None, bias_initializer='zeros',
in_channels=0, **kwargs):
+ assert layout == 'NCW', "Only supports 'NCW' layout for now"
if isinstance(kernel_size, numeric_types):
kernel_size = (kernel_size,)
if isinstance(output_padding, numeric_types):
@@ -513,9 +519,9 @@ class Conv2DTranspose(_Conv):
layers side by side, each seeing half the input channels, and producing
half the output channels, and both subsequently concatenated.
layout : str, default 'NCHW'
- Dimension ordering of data and weight. Can be 'NCHW', 'NHWC', etc.
- 'N', 'C', 'H', 'W' stands for batch, channel, height, and width
- dimensions respectively. Convolution is applied on the 'H' and
+ Dimension ordering of data and weight. Only supports 'NCHW' and 'NHWC'
+ layout for now. 'N', 'C', 'H', 'W' stands for batch, channel, height,
+ and width dimensions respectively. Convolution is applied on the 'H' and
'W' dimensions.
in_channels : int, default 0
The number of input channels to this layer. If not specified,
@@ -535,12 +541,12 @@ class Conv2DTranspose(_Conv):
Inputs:
- **data**: 4D input tensor with shape
- `(batch_size, in_channels, height, width)` when `layout` is `NCW`.
+ `(batch_size, in_channels, height, width)` when `layout` is `NCHW`.
For other layouts shape is permuted accordingly.
Outputs:
- **out**: 4D output tensor with shape
- `(batch_size, channels, out_height, out_width)` when `layout` is `NCW`.
+ `(batch_size, channels, out_height, out_width)` when `layout` is `NCHW`.
out_height and out_width are calculated as::
out_height = (height-1)*strides[0]-2*padding[0]+kernel_size[0]+output_padding[0]
@@ -550,6 +556,8 @@ class Conv2DTranspose(_Conv):
output_padding=(0, 0), dilation=(1, 1), groups=1, layout='NCHW',
activation=None, use_bias=True, weight_initializer=None,
bias_initializer='zeros', in_channels=0, **kwargs):
+ assert layout == 'NCHW' or layout == 'NHWC', \
+ "Only supports 'NCHW' and 'NHWC' layout for now"
if isinstance(kernel_size, numeric_types):
kernel_size = (kernel_size,)*2
if isinstance(output_padding, numeric_types):
@@ -599,10 +607,10 @@ class Conv3DTranspose(_Conv):
layers side by side, each seeing half the input channels, and producing
half the output channels, and both subsequently concatenated.
layout : str, default 'NCDHW'
- Dimension ordering of data and weight. Can be 'NCDHW', 'NDHWC', etc.
- 'N', 'C', 'H', 'W', 'D' stands for batch, channel, height, width and
- depth dimensions respectively. Convolution is applied on the 'D',
- 'H', and 'W' dimensions.
+ Dimension ordering of data and weight. Only supports 'NCDHW' and 'NDHWC'
+ layout for now. 'N', 'C', 'H', 'W', 'D' stands for batch, channel, height,
+ width and depth dimensions respectively. Convolution is applied on the 'D',
+ 'H' and 'W' dimensions.
in_channels : int, default 0
The number of input channels to this layer. If not specified,
initialization will be deferred to the first time `forward` is called
@@ -621,12 +629,12 @@ class Conv3DTranspose(_Conv):
Inputs:
- **data**: 5D input tensor with shape
- `(batch_size, in_channels, depth, height, width)` when `layout` is `NCW`.
+ `(batch_size, in_channels, depth, height, width)` when `layout` is `NCDHW`.
For other layouts shape is permuted accordingly.
Outputs:
- **out**: 5D output tensor with shape
- `(batch_size, channels, out_depth, out_height, out_width)` when `layout` is `NCW`.
+ `(batch_size, channels, out_depth, out_height, out_width)` when `layout` is `NCDHW`.
out_depth, out_height and out_width are calculated as::
out_depth = (depth-1)*strides[0]-2*padding[0]+kernel_size[0]+output_padding[0]
@@ -637,6 +645,8 @@ class Conv3DTranspose(_Conv):
output_padding=(0, 0, 0), dilation=(1, 1, 1), groups=1, layout='NCDHW',
activation=None, use_bias=True, weight_initializer=None,
bias_initializer='zeros', in_channels=0, **kwargs):
+ assert layout == 'NCDHW' or layout == 'NDHWC', \
+ "Only supports 'NCDHW' and 'NDHWC' layout for now"
if isinstance(kernel_size, numeric_types):
kernel_size = (kernel_size,)*3
if isinstance(output_padding, numeric_types):
@@ -694,7 +704,7 @@ class MaxPool1D(_Pooling):
If padding is non-zero, then the input is implicitly
zero-padded on both sides for padding number of points.
layout : str, default 'NCW'
- Dimension ordering of data and weight. Can be 'NCW', 'NWC', etc.
+ Dimension ordering of data and weight. Only supports 'NCW' layout for now.
'N', 'C', 'W' stands for batch, channel, and width (time) dimensions
respectively. Pooling is applied on the W dimension.
ceil_mode : bool, default False
@@ -716,7 +726,7 @@ class MaxPool1D(_Pooling):
"""
def __init__(self, pool_size=2, strides=None, padding=0, layout='NCW',
ceil_mode=False, **kwargs):
- assert layout == 'NCW', "Only supports NCW layout for now"
+ assert layout == 'NCW', "Only supports 'NCW' layout for now"
if isinstance(pool_size, numeric_types):
pool_size = (pool_size,)
assert len(pool_size) == 1, "pool_size must be a number or a list of 1 ints"
@@ -739,7 +749,7 @@ class MaxPool2D(_Pooling):
If padding is non-zero, then the input is implicitly
zero-padded on both sides for padding number of points.
layout : str, default 'NCHW'
- Dimension ordering of data and weight. Can be 'NCHW', 'NHWC', etc.
+ Dimension ordering of data and weight. Only supports 'NCHW' layout for now.
'N', 'C', 'H', 'W' stands for batch, channel, height, and width
dimensions respectively. padding is applied on 'H' and 'W' dimension.
ceil_mode : bool, default False
@@ -748,12 +758,12 @@ class MaxPool2D(_Pooling):
Inputs:
- **data**: 4D input tensor with shape
- `(batch_size, in_channels, height, width)` when `layout` is `NCW`.
+ `(batch_size, in_channels, height, width)` when `layout` is `NCHW`.
For other layouts shape is permuted accordingly.
Outputs:
- **out**: 4D output tensor with shape
- `(batch_size, channels, out_height, out_width)` when `layout` is `NCW`.
+ `(batch_size, channels, out_height, out_width)` when `layout` is `NCHW`.
out_height and out_width are calculated as::
out_height = floor((height+2*padding[0]-pool_size[0])/strides[0])+1
@@ -764,7 +774,7 @@ class MaxPool2D(_Pooling):
"""
def __init__(self, pool_size=(2, 2), strides=None, padding=0, layout='NCHW',
ceil_mode=False, **kwargs):
- assert layout == 'NCHW', "Only supports NCHW layout for now"
+ assert layout == 'NCHW', "Only supports 'NCHW' layout for now"
if isinstance(pool_size, numeric_types):
pool_size = (pool_size,)*2
assert len(pool_size) == 2, "pool_size must be a number or a list of 2 ints"
@@ -787,7 +797,7 @@ class MaxPool3D(_Pooling):
If padding is non-zero, then the input is implicitly
zero-padded on both sides for padding number of points.
layout : str, default 'NCDHW'
- Dimension ordering of data and weight. Can be 'NCDHW', 'NDHWC', etc.
+ Dimension ordering of data and weight. Only supports 'NCDHW' layout for now.
'N', 'C', 'H', 'W', 'D' stands for batch, channel, height, width and
depth dimensions respectively. padding is applied on 'D', 'H' and 'W'
dimension.
@@ -802,7 +812,7 @@ class MaxPool3D(_Pooling):
Outputs:
- **out**: 5D output tensor with shape
- `(batch_size, channels, out_depth, out_height, out_width)` when `layout` is `NCW`.
+ `(batch_size, channels, out_depth, out_height, out_width)` when `layout` is `NCDHW`.
out_depth, out_height and out_width are calculated as::
out_depth = floor((depth+2*padding[0]-pool_size[0])/strides[0])+1
@@ -814,7 +824,7 @@ class MaxPool3D(_Pooling):
"""
def __init__(self, pool_size=(2, 2, 2), strides=None, padding=0,
ceil_mode=False, layout='NCDHW', **kwargs):
- assert layout == 'NCDHW', "Only supports NCDHW layout for now"
+ assert layout == 'NCDHW', "Only supports 'NCDHW' layout for now"
if isinstance(pool_size, numeric_types):
pool_size = (pool_size,)*3
assert len(pool_size) == 3, "pool_size must be a number or a list of 3 ints"
@@ -836,7 +846,7 @@ class AvgPool1D(_Pooling):
If padding is non-zero, then the input is implicitly
zero-padded on both sides for padding number of points.
layout : str, default 'NCW'
- Dimension ordering of data and weight. Can be 'NCW', 'NWC', etc.
+ Dimension ordering of data and weight. Only supports 'NCW' layout for now.
'N', 'C', 'W' stands for batch, channel, and width (time) dimensions
respectively. padding is applied on 'W' dimension.
ceil_mode : bool, default False
@@ -858,7 +868,7 @@ class AvgPool1D(_Pooling):
"""
def __init__(self, pool_size=2, strides=None, padding=0, layout='NCW',
ceil_mode=False, **kwargs):
- assert layout == 'NCW', "Only supports NCW layout for now"
+ assert layout == 'NCW', "Only supports 'NCW' layout for now"
if isinstance(pool_size, numeric_types):
pool_size = (pool_size,)
assert len(pool_size) == 1, "pool_size must be a number or a list of 1 ints"
@@ -880,7 +890,7 @@ class AvgPool2D(_Pooling):
If padding is non-zero, then the input is implicitly
zero-padded on both sides for padding number of points.
layout : str, default 'NCHW'
- Dimension ordering of data and weight. Can be 'NCHW', 'NHWC', etc.
+ Dimension ordering of data and weight. Only supports 'NCHW' layout for now.
'N', 'C', 'H', 'W' stands for batch, channel, height, and width
dimensions respectively. padding is applied on 'H' and 'W' dimension.
ceil_mode : bool, default False
@@ -889,12 +899,12 @@ class AvgPool2D(_Pooling):
Inputs:
- **data**: 4D input tensor with shape
- `(batch_size, in_channels, height, width)` when `layout` is `NCW`.
+ `(batch_size, in_channels, height, width)` when `layout` is `NCHW`.
For other layouts shape is permuted accordingly.
Outputs:
- **out**: 4D output tensor with shape
- `(batch_size, channels, out_height, out_width)` when `layout` is `NCW`.
+ `(batch_size, channels, out_height, out_width)` when `layout` is `NCHW`.
out_height and out_width are calculated as::
out_height = floor((height+2*padding[0]-pool_size[0])/strides[0])+1
@@ -905,7 +915,7 @@ class AvgPool2D(_Pooling):
"""
def __init__(self, pool_size=(2, 2), strides=None, padding=0,
ceil_mode=False, layout='NCHW', **kwargs):
- assert layout == 'NCHW', "Only supports NCHW layout for now"
+ assert layout == 'NCHW', "Only supports 'NCHW' layout for now"
if isinstance(pool_size, numeric_types):
pool_size = (pool_size,)*2
assert len(pool_size) == 2, "pool_size must be a number or a list of 2 ints"
@@ -937,12 +947,12 @@ class AvgPool3D(_Pooling):
Inputs:
- **data**: 5D input tensor with shape
- `(batch_size, in_channels, depth, height, width)` when `layout` is `NCW`.
+ `(batch_size, in_channels, depth, height, width)` when `layout` is `NCDHW`.
For other layouts shape is permuted accordingly.
Outputs:
- **out**: 5D output tensor with shape
- `(batch_size, channels, out_depth, out_height, out_width)` when `layout` is `NCW`.
+ `(batch_size, channels, out_depth, out_height, out_width)` when `layout` is `NCDHW`.
out_depth, out_height and out_width are calculated as::
out_depth = floor((depth+2*padding[0]-pool_size[0])/strides[0])+1
@@ -954,7 +964,7 @@ class AvgPool3D(_Pooling):
"""
def __init__(self, pool_size=(2, 2, 2), strides=None, padding=0,
ceil_mode=False, layout='NCDHW', **kwargs):
- assert layout == 'NCDHW', "Only supports NCDHW layout for now"
+ assert layout == 'NCDHW', "Only supports 'NCDHW' layout for now"
if isinstance(pool_size, numeric_types):
pool_size = (pool_size,)*3
assert len(pool_size) == 3, "pool_size must be a number or a list of 3 ints"
@@ -965,7 +975,7 @@ class AvgPool3D(_Pooling):
class GlobalMaxPool1D(_Pooling):
"""Global max pooling operation for temporal data."""
def __init__(self, layout='NCW', **kwargs):
- assert layout == 'NCW', "Only supports NCW layout for now"
+ assert layout == 'NCW', "Only supports 'NCW' layout for now"
super(GlobalMaxPool1D, self).__init__(
(1,), None, 0, True, True, 'max', **kwargs)
@@ -973,7 +983,7 @@ class GlobalMaxPool1D(_Pooling):
class GlobalMaxPool2D(_Pooling):
"""Global max pooling operation for spatial data."""
def __init__(self, layout='NCHW', **kwargs):
- assert layout == 'NCHW', "Only supports NCHW layout for now"
+ assert layout == 'NCHW', "Only supports 'NCHW' layout for now"
super(GlobalMaxPool2D, self).__init__(
(1, 1), None, 0, True, True, 'max', **kwargs)
@@ -981,7 +991,7 @@ class GlobalMaxPool2D(_Pooling):
class GlobalMaxPool3D(_Pooling):
"""Global max pooling operation for 3D data."""
def __init__(self, layout='NCDHW', **kwargs):
- assert layout == 'NCDHW', "Only supports NCDHW layout for now"
+ assert layout == 'NCDHW', "Only supports 'NCDHW' layout for now"
super(GlobalMaxPool3D, self).__init__(
(1, 1, 1), None, 0, True, True, 'max', **kwargs)
@@ -989,7 +999,7 @@ class GlobalMaxPool3D(_Pooling):
class GlobalAvgPool1D(_Pooling):
"""Global average pooling operation for temporal data."""
def __init__(self, layout='NCW', **kwargs):
- assert layout == 'NCW', "Only supports NCW layout for now"
+ assert layout == 'NCW', "Only supports 'NCW' layout for now"
super(GlobalAvgPool1D, self).__init__(
(1,), None, 0, True, True, 'avg', **kwargs)
@@ -997,7 +1007,7 @@ class GlobalAvgPool1D(_Pooling):
class GlobalAvgPool2D(_Pooling):
"""Global average pooling operation for spatial data."""
def __init__(self, layout='NCHW', **kwargs):
- assert layout == 'NCHW', "Only supports NCHW layout for now"
+ assert layout == 'NCHW', "Only supports 'NCHW' layout for now"
super(GlobalAvgPool2D, self).__init__(
(1, 1), None, 0, True, True, 'avg', **kwargs)
@@ -1005,7 +1015,7 @@ class GlobalAvgPool2D(_Pooling):
class GlobalAvgPool3D(_Pooling):
"""Global max pooling operation for 3D data."""
def __init__(self, layout='NCDHW', **kwargs):
- assert layout == 'NCDHW', "Only supports NCDHW layout for now"
+ assert layout == 'NCDHW', "Only supports 'NCDHW' layout for now"
super(GlobalAvgPool3D, self).__init__(
(1, 1, 1), None, 0, True, True, 'avg', **kwargs)
--
To stop receiving notification emails like this one, please contact
jxie@apache.org.