File tree Expand file tree Collapse file tree 1 file changed +68
-25
lines changed Expand file tree Collapse file tree 1 file changed +68
-25
lines changed Original file line number Diff line number Diff line change @@ -5,54 +5,97 @@ Operators
55
66.. currentmodule :: torchvision.ops
77
8- :mod: `torchvision.ops ` implements operators that are specific for Computer Vision.
8+ :mod: `torchvision.ops ` implements operators, losses and layers that are specific for Computer Vision.
99
1010.. note ::
1111 All operators have native support for TorchScript.
1212
13+
14+ Detection and Segmentation Operators
15+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
16+
17+ The below operators perform pre-processing as well as post-processing required in object detection and segmentation models.
18+
1319.. autosummary ::
1420 :toctree: generated/
1521 :template: function.rst
1622
1723 batched_nms
24+ masks_to_boxes
25+ nms
26+ roi_align
27+ roi_pool
28+ ps_roi_align
29+ ps_roi_pool
30+
31+ .. autosummary ::
32+ :toctree: generated/
33+ :template: class.rst
34+
35+ FeaturePyramidNetwork
36+ MultiScaleRoIAlign
37+ RoIAlign
38+ RoIPool
39+ PSRoIAlign
40+ PSRoIPool
41+
42+
43+ Box Operators
44+ ~~~~~~~~~~~~~
45+
46+ These utility functions perform various operations on bounding boxes.
47+
48+ .. autosummary ::
49+ :toctree: generated/
50+ :template: function.rst
51+
1852 box_area
1953 box_convert
2054 box_iou
2155 clip_boxes_to_image
2256 complete_box_iou
23- complete_box_iou_loss
24- deform_conv2d
25- drop_block2d
26- drop_block3d
27- generalized_box_iou
28- generalized_box_iou_loss
2957 distance_box_iou
30- distance_box_iou_loss
31- masks_to_boxes
32- nms
33- ps_roi_align
34- ps_roi_pool
58+ generalized_box_iou
3559 remove_small_boxes
36- roi_align
37- roi_pool
60+
61+ Losses
62+ ~~~~~~
63+
64+ The following vision-specific loss functions are implemented:
65+
66+ .. autosummary ::
67+ :toctree: generated/
68+ :template: function.rst
69+
70+ complete_box_iou_loss
71+ distance_box_iou_loss
72+ generalized_box_iou_loss
3873 sigmoid_focal_loss
39- stochastic_depth
74+
75+
76+ Layers
77+ ~~~~~~
78+
79+ TorchVision provides commonly used building blocks as layers:
4080
4181.. autosummary ::
4282 :toctree: generated/
4383 :template: class.rst
4484
45- RoIAlign
46- PSRoIAlign
47- RoIPool
48- PSRoIPool
49- DeformConv2d
50- MultiScaleRoIAlign
51- FeaturePyramidNetwork
52- StochasticDepth
53- FrozenBatchNorm2d
5485 Conv2dNormActivation
5586 Conv3dNormActivation
56- SqueezeExcitation
87+ DeformConv2d
5788 DropBlock2d
5889 DropBlock3d
90+ FrozenBatchNorm2d
91+ SqueezeExcitation
92+ StochasticDepth
93+
94+ .. autosummary ::
95+ :toctree: generated/
96+ :template: function.rst
97+
98+ deform_conv2d
99+ drop_block2d
100+ drop_block3d
101+ stochastic_depth
You can’t perform that action at this time.
0 commit comments