.. _op_ai_onnx_Resize:
Resize
======
- **Domain**: ``ai.onnx``
- **Since version**: 19
Resize the input tensor. In general, it calculates every value in the output tensor as a weighted average of neighborhood (a.k.a. sampling locations) in the input tensor.
Each dimension value of the output tensor is:
.. code-block::
output_dimension = floor(input_dimension * (roi_end - roi_start) * scale)
if input \"sizes\" is not specified.
**Inputs**
- **X** (*T1*): N-D tensor
- **roi** (*T2*): 1-D tensor given as [start1, ..., startN, end1, ..., endN], where N is the rank of X or the length of axes, if provided. The RoIs' coordinates are normalized in the coordinate system of the input image. It only takes effect when coordinate_transformation_mode is "tf_crop_and_resize"
- **scales** (*tensor(float)*): The scale array along each dimension. It takes value greater than 0. If it's less than 1, it's sampling down, otherwise, it's upsampling. The number of elements of 'scales' should be the same as the rank of input 'X' or the length of 'axes', if provided. One of 'scales' and 'sizes' MUST be specified and it is an error if both are specified. If 'sizes' is needed, the user can use an empty string as the name of 'scales' in this operator's input list.
- **sizes** (*tensor(int64)*): Target size of the output tensor. Its interpretation depends on the 'keep_aspect_ratio_policy' value.The number of elements of 'sizes' should be the same as the rank of input 'X', or the length of 'axes', if provided. Only one of 'scales' and 'sizes' can be specified.
**Outputs**
- **Y** (*T1*): N-D tensor after resizing
**Type Constraints**
- **T1**: Constrain input 'X' and output 'Y' to all tensor types.
Allowed types: tensor(bfloat16), tensor(bool), tensor(complex128), tensor(complex64), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(string), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8).
- **T2**: Constrain roi type to float or double.
Allowed types: tensor(double), tensor(float), tensor(float16).
Examples
--------
**test_resize_downsample_scales_cubic**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
mode = "cubic"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
scales: shape=(4,), dtype=float32
[1. , 1. , 0.8, 0.8]
Outputs:
Y: shape=(1, 1, 3, 3), dtype=float32
[[[[ 1.4711914, 2.78125 , 4.0825195],
[ 6.711426 , 8.021484 , 9.322754 ],
[11.916504 , 13.2265625, 14.527832 ]]]]
**test_resize_downsample_scales_cubic_A_n0p5_exclude_outside**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
cubic_coeff_a = -0.5
exclude_outside = 1
mode = "cubic"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
scales: shape=(4,), dtype=float32
[1. , 1. , 0.8, 0.8]
Outputs:
Y: shape=(1, 1, 3, 3), dtype=float32
[[[[ 1.3681267, 2.6695013, 4.0133367],
[ 6.573625 , 7.875 , 9.218835 ],
[11.948966 , 13.250341 , 14.594176 ]]]]
**test_resize_downsample_scales_cubic_align_corners**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
coordinate_transformation_mode = "align_corners"
mode = "cubic"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
scales: shape=(4,), dtype=float32
[1. , 1. , 0.8, 0.8]
Outputs:
Y: shape=(1, 1, 3, 3), dtype=float32
[[[[ 1. , 2.3951917, 3.790383 ],
[ 6.580766 , 7.975958 , 9.371149 ],
[12.161532 , 13.556724 , 14.951916 ]]]]
**test_resize_downsample_scales_cubic_antialias**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
antialias = 1
mode = "cubic"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
scales: shape=(4,), dtype=float32
[1. , 1. , 0.6, 0.6]
Outputs:
Y: shape=(1, 1, 2, 2), dtype=float32
[[[[ 2.5180721, 4.2858863],
[ 9.589329 , 11.357142 ]]]]
**test_resize_downsample_scales_linear**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 2, 4), dtype=float32
[[[[1., 2., 3., 4.],
[5., 6., 7., 8.]]]]
scales: shape=(4,), dtype=float32
[1. , 1. , 0.6, 0.6]
Outputs:
Y: shape=(1, 1, 1, 2), dtype=float32
[[[[2.6666665, 4.333333 ]]]]
**test_resize_downsample_scales_linear_align_corners**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
coordinate_transformation_mode = "align_corners"
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 2, 4), dtype=float32
[[[[1., 2., 3., 4.],
[5., 6., 7., 8.]]]]
scales: shape=(4,), dtype=float32
[1. , 1. , 0.6, 0.6]
Outputs:
Y: shape=(1, 1, 1, 2), dtype=float32
[[[[1. , 3.142857]]]]
**test_resize_downsample_scales_linear_antialias**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
antialias = 1
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
scales: shape=(4,), dtype=float32
[1. , 1. , 0.6, 0.6]
Outputs:
Y: shape=(1, 1, 2, 2), dtype=float32
[[[[ 2.875, 4.5 ],
[ 9.375, 11. ]]]]
**test_resize_downsample_scales_linear_half_pixel_symmetric**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
coordinate_transformation_mode = "half_pixel_symmetric"
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 1, 4), dtype=float32
[[[[1., 2., 3., 4.]]]]
scales: shape=(4,), dtype=float32
[1. , 1. , 1. , 0.6]
Outputs:
Y: shape=(1, 1, 1, 2), dtype=float32
[[[[1.6666667, 3.3333333]]]]
**test_resize_downsample_sizes_cubic**
.. code-block:: text
Node:
Resize(X, "", "", sizes) -> (Y)
Attributes:
mode = "cubic"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
sizes: shape=(4,), dtype=int64
[1, 1, 3, 3]
Outputs:
Y: shape=(1, 1, 3, 3), dtype=float32
[[[[ 1.630787 , 3.0046296, 4.3784723],
[ 7.1261573, 8.5 , 9.873842 ],
[12.621528 , 13.99537 , 15.369213 ]]]]
**test_resize_downsample_sizes_cubic_antialias**
.. code-block:: text
Node:
Resize(X, "", "", sizes) -> (Y)
Attributes:
antialias = 1
mode = "cubic"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
sizes: shape=(4,), dtype=int64
[1, 1, 3, 3]
Outputs:
Y: shape=(1, 1, 3, 3), dtype=float32
[[[[ 1.7750092, 3.1200073, 4.4650054],
[ 7.1550016, 8.5 , 9.844998 ],
[12.534994 , 13.8799925, 15.224991 ]]]]
**test_resize_downsample_sizes_linear_antialias**
.. code-block:: text
Node:
Resize(X, "", "", sizes) -> (Y)
Attributes:
antialias = 1
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
sizes: shape=(4,), dtype=int64
[1, 1, 3, 3]
Outputs:
Y: shape=(1, 1, 3, 3), dtype=float32
[[[[ 2.3636363, 3.590909 , 4.818182 ],
[ 7.2727275, 8.5 , 9.727273 ],
[12.181818 , 13.409091 , 14.636364 ]]]]
**test_resize_downsample_sizes_linear_pytorch_half_pixel**
.. code-block:: text
Node:
Resize(X, "", "", sizes) -> (Y)
Attributes:
coordinate_transformation_mode = "pytorch_half_pixel"
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
sizes: shape=(4,), dtype=int64
[1, 1, 3, 1]
Outputs:
Y: shape=(1, 1, 3, 1), dtype=float32
[[[[ 1.6666666],
[ 7. ],
[12.333333 ]]]]
**test_resize_tf_crop_and_resize**
.. code-block:: text
Node:
Resize(X, roi, "", sizes) -> (Y)
Attributes:
coordinate_transformation_mode = "tf_crop_and_resize"
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
roi: shape=(8,), dtype=float32
[0. , 0. , 0.4, 0.6, 1. , 1. , 0.6, 0.8]
sizes: shape=(4,), dtype=int64
[1, 1, 3, 3]
Outputs:
Y: shape=(1, 1, 3, 3), dtype=float32
[[[[ 7.6000004, 7.9 , 8.2 ],
[ 8.8 , 9.1 , 9.400001 ],
[10. , 10.3 , 10.6 ]]]]
**test_resize_tf_crop_and_resize_axes_2_3**
.. code-block:: text
Node:
Resize(X, roi, "", sizes) -> (Y)
Attributes:
axes = [2, 3]
coordinate_transformation_mode = "tf_crop_and_resize"
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
roi: shape=(4,), dtype=float32
[0.4, 0.6, 0.6, 0.8]
sizes: shape=(2,), dtype=int64
[3, 3]
Outputs:
Y: shape=(1, 1, 3, 3), dtype=float32
[[[[ 7.6000004, 7.9 , 8.2 ],
[ 8.8 , 9.1 , 9.400001 ],
[10. , 10.3 , 10.6 ]]]]
**test_resize_tf_crop_and_resize_axes_3_2**
.. code-block:: text
Node:
Resize(X, roi, "", sizes) -> (Y)
Attributes:
axes = [3, 2]
coordinate_transformation_mode = "tf_crop_and_resize"
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
roi: shape=(4,), dtype=float32
[0.6, 0.4, 0.8, 0.6]
sizes: shape=(2,), dtype=int64
[3, 3]
Outputs:
Y: shape=(1, 1, 3, 3), dtype=float32
[[[[ 7.6000004, 7.9 , 8.2 ],
[ 8.8 , 9.1 , 9.400001 ],
[10. , 10.3 , 10.6 ]]]]
**test_resize_tf_crop_and_resize_extrapolation_value**
.. code-block:: text
Node:
Resize(X, roi, "", sizes) -> (Y)
Attributes:
coordinate_transformation_mode = "tf_crop_and_resize"
extrapolation_value = 10.0
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
roi: shape=(8,), dtype=float32
[0. , 0. , 0.4, 0.6, 1. , 1. , 1.2, 1.7]
sizes: shape=(4,), dtype=int64
[1, 1, 3, 3]
Outputs:
Y: shape=(1, 1, 3, 3), dtype=float32
[[[[ 7.6000004, 10. , 10. ],
[12.400001 , 10. , 10. ],
[10. , 10. , 10. ]]]]
**test_resize_upsample_scales_cubic**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
mode = "cubic"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
scales: shape=(4,), dtype=float32
[1., 1., 2., 2.]
Outputs:
Y: shape=(1, 1, 8, 8), dtype=float32
[[[[ 0.47265625, 0.76953125, 1.2460938 , 1.875 , 2.28125 ,
2.9101562 , 3.3867188 , 3.6835938 ],
[ 1.6601562 , 1.9570312 , 2.4335938 , 3.0625 , 3.46875 ,
4.0976562 , 4.5742188 , 4.8710938 ],
[ 3.5664062 , 3.8632812 , 4.3398438 , 4.96875 , 5.375 ,
6.0039062 , 6.4804688 , 6.7773438 ],
[ 6.0820312 , 6.3789062 , 6.8554688 , 7.484375 , 7.890625 ,
8.519531 , 8.996094 , 9.292969 ],
[ 7.7070312 , 8.003906 , 8.480469 , 9.109375 , 9.515625 ,
10.144531 , 10.621094 , 10.917969 ],
[10.222656 , 10.519531 , 10.996094 , 11.625 , 12.03125 ,
12.660156 , 13.136719 , 13.433594 ],
[12.128906 , 12.425781 , 12.902344 , 13.53125 , 13.9375 ,
14.566406 , 15.042969 , 15.339844 ],
[13.316406 , 13.613281 , 14.089844 , 14.71875 , 15.125 ,
15.753906 , 16.230469 , 16.527344 ]]]]
**test_resize_upsample_scales_cubic_A_n0p5_exclude_outside**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
cubic_coeff_a = -0.5
exclude_outside = 1
mode = "cubic"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
scales: shape=(4,), dtype=float32
[1., 1., 2., 2.]
Outputs:
Y: shape=(1, 1, 8, 8), dtype=float32
[[[[ 0.5588235 , 0.81494206, 1.3569825 , 1.8970588 , 2.3970587 ,
2.9371352 , 3.4791756 , 3.735294 ],
[ 1.5832976 , 1.839416 , 2.3814566 , 2.9215329 , 3.4215329 ,
3.9616091 , 4.5036497 , 4.759768 ],
[ 3.7514594 , 4.007578 , 4.5496182 , 5.0896945 , 5.5896945 ,
6.1297708 , 6.6718116 , 6.92793 ],
[ 5.9117646 , 6.1678834 , 6.7099237 , 7.25 , 7.75 ,
8.290076 , 8.832117 , 9.088235 ],
[ 7.9117646 , 8.167883 , 8.709924 , 9.25 , 9.75 ,
10.290076 , 10.832117 , 11.088235 ],
[10.07207 , 10.328189 , 10.870229 , 11.410305 , 11.910305 ,
12.450381 , 12.992422 , 13.248541 ],
[12.2402315 , 12.49635 , 13.038391 , 13.578467 , 14.078467 ,
14.618544 , 15.1605835 , 15.416702 ],
[13.264706 , 13.520824 , 14.062865 , 14.6029415 , 15.1029415 ,
15.643018 , 16.185059 , 16.441177 ]]]]
**test_resize_upsample_scales_cubic_align_corners**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
coordinate_transformation_mode = "align_corners"
mode = "cubic"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
scales: shape=(4,), dtype=float32
[1., 1., 2., 2.]
Outputs:
Y: shape=(1, 1, 8, 8), dtype=float32
[[[[ 1. , 1.3411078, 1.8002915, 2.329446 , 2.670554 , 3.1997085,
3.6588922, 4. ],
[ 2.3644314, 2.7055395, 3.164723 , 3.6938775, 4.0349855, 4.56414 ,
5.0233235, 5.3644314],
[ 4.201166 , 4.542274 , 5.0014577, 5.5306125, 5.8717203, 6.4008746,
6.8600583, 7.201166 ],
[ 6.3177843, 6.658892 , 7.118076 , 7.64723 , 7.988338 , 8.517492 ,
8.976676 , 9.317784 ],
[ 7.6822157, 8.023324 , 8.482508 , 9.011662 , 9.35277 , 9.881925 ,
10.341108 , 10.682216 ],
[ 9.798834 , 10.139941 , 10.599125 , 11.12828 , 11.469388 , 11.998542 ,
12.457726 , 12.798834 ],
[11.635569 , 11.976676 , 12.43586 , 12.965014 , 13.306123 , 13.835277 ,
14.29446 , 14.635569 ],
[13. , 13.341108 , 13.800292 , 14.329446 , 14.670554 , 15.199708 ,
15.658892 , 16. ]]]]
**test_resize_upsample_scales_cubic_asymmetric**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
coordinate_transformation_mode = "asymmetric"
mode = "cubic"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
scales: shape=(4,), dtype=float32
[1., 1., 2., 2.]
Outputs:
Y: shape=(1, 1, 8, 8), dtype=float32
[[[[ 1. , 1.40625, 2. , 2.5 , 3. , 3.59375, 4. ,
4.09375],
[ 2.625 , 3.03125, 3.625 , 4.125 , 4.625 , 5.21875, 5.625 ,
5.71875],
[ 5. , 5.40625, 6. , 6.5 , 7. , 7.59375, 8. ,
8.09375],
[ 7. , 7.40625, 8. , 8.5 , 9. , 9.59375, 10. ,
10.09375],
[ 9. , 9.40625, 10. , 10.5 , 11. , 11.59375, 12. ,
12.09375],
[11.375 , 11.78125, 12.375 , 12.875 , 13.375 , 13.96875, 14.375 ,
14.46875],
[13. , 13.40625, 14. , 14.5 , 15. , 15.59375, 16. ,
16.09375],
[13.375 , 13.78125, 14.375 , 14.875 , 15.375 , 15.96875, 16.375 ,
16.46875]]]]
**test_resize_upsample_scales_linear**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 2, 2), dtype=float32
[[[[1., 2.],
[3., 4.]]]]
scales: shape=(4,), dtype=float32
[1., 1., 2., 2.]
Outputs:
Y: shape=(1, 1, 4, 4), dtype=float32
[[[[1. , 1.25, 1.75, 2. ],
[1.5 , 1.75, 2.25, 2.5 ],
[2.5 , 2.75, 3.25, 3.5 ],
[3. , 3.25, 3.75, 4. ]]]]
**test_resize_upsample_scales_linear_align_corners**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
coordinate_transformation_mode = "align_corners"
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 2, 2), dtype=float32
[[[[1., 2.],
[3., 4.]]]]
scales: shape=(4,), dtype=float32
[1., 1., 2., 2.]
Outputs:
Y: shape=(1, 1, 4, 4), dtype=float32
[[[[1. , 1.3333334, 1.6666666, 2. ],
[1.6666666, 2. , 2.3333333, 2.6666667],
[2.3333333, 2.6666667, 3. , 3.3333333],
[3. , 3.3333333, 3.6666667, 4. ]]]]
**test_resize_upsample_scales_linear_half_pixel_symmetric**
.. code-block:: text
Node:
Resize(X, "", scales) -> (Y)
Attributes:
coordinate_transformation_mode = "half_pixel_symmetric"
mode = "linear"
.. code-block:: text
Inputs:
X: shape=(1, 1, 2, 2), dtype=float32
[[[[1., 2.],
[3., 4.]]]]
scales: shape=(4,), dtype=float32
[1. , 1. , 2.3 , 2.94]
Outputs:
Y: shape=(1, 1, 4, 5), dtype=float32
[[[[1. , 1.159864 , 1.5 , 1.840136 , 2. ],
[1.5652174, 1.7250813, 2.0652175, 2.4053535, 2.5652175],
[2.4347825, 2.5946465, 2.9347825, 3.2749186, 3.4347825],
[3. , 3.159864 , 3.5 , 3.840136 , 4. ]]]]
**test_resize_upsample_sizes_cubic**
.. code-block:: text
Node:
Resize(X, "", "", sizes) -> (Y)
Attributes:
mode = "cubic"
.. code-block:: text
Inputs:
X: shape=(1, 1, 4, 4), dtype=float32
[[[[ 1., 2., 3., 4.],
[ 5., 6., 7., 8.],
[ 9., 10., 11., 12.],
[13., 14., 15., 16.]]]]
sizes: shape=(4,), dtype=int64
[ 1, 1, 9, 10]
Outputs:
Y: shape=(1, 1, 9, 10), dtype=float32
[[[[ 0.45507923, 0.6405792 , 0.9715792 , ..., 3.1590793 , 3.4900792 ,
3.6755793 ],
[ 1.3943796 , 1.5798796 , 1.9108796 , ..., 4.0983796 , 4.4293795 ,
4.6148796 ],
[ 2.9513068 , 3.136807 , 3.4678068 , ..., 5.655307 , 5.986307 ,
6.171807 ],
...,
[10.828193 , 11.013693 , 11.344693 , ..., 13.532193 , 13.8631935 ,
14.048693 ],
[12.38512 , 12.570621 , 12.90162 , ..., 15.08912 , 15.42012 ,
15.60562 ],
[13.324421 , 13.509921 , 13.84092 , ..., 16.028421 , 16.35942 ,
16.54492 ]]]]
Differences with previous version (18)
--------------------------------------
**SchemaDiff**: ``Resize`` (domain ``'ai.onnx'``)
* old version: 18
* new version: 19
* breaking: no
**Documentation:**
* line similarity: 0.50 (+4/-2 lines)
.. code-block:: diff
--- Resize v18
+++ Resize v19
@@ -1,5 +1,7 @@
Resize the input tensor. In general, it calculates every value in the output tensor as a weighted average of neighborhood (a.k.a. sampling locations) in the input tensor.
-Each dimension value of the output tensor is:
- `output_dimension = floor(input_dimension * (roi_end - roi_start) * scale)`
+Each dimension value of the output tensor is:
+```
+output_dimension = floor(input_dimension * (roi_end - roi_start) * scale)
+```
if input \"sizes\" is not specified.
Version History
---------------
- :doc:`Version 18 `
- :doc:`Version 13 `
- :doc:`Version 11 `
- :doc:`Version 10 `