models.rst 26.5 KB
Newer Older
1
2
.. _models:

3
4
Models and pre-trained weights
##############################
5
6


7
The ``torchvision.models`` subpackage contains definitions of models for addressing
8
different tasks, including: image classification, pixelwise semantic
9
segmentation, object detection, instance segmentation, person
10
keypoint detection, video classification, and optical flow.
11

12
13
.. note ::
    Backward compatibility is guaranteed for loading a serialized 
14
    ``state_dict`` to the model created using old PyTorch version. 
15
    On the contrary, loading entire saved models or serialized 
16
17
    ``ScriptModules`` (seralized using older versions of PyTorch) 
    may not preserve the historic behaviour. Refer to the following 
18
19
20
    `documentation 
    <https://pytorch.org/docs/stable/notes/serialization.html#id6>`_   

21
22
23

Classification
==============
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
24

Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
25
The models subpackage contains definitions for the following model
26
architectures for image classification:
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
27
28
29
30
31
32
33

-  `AlexNet`_
-  `VGG`_
-  `ResNet`_
-  `SqueezeNet`_
-  `DenseNet`_
-  `Inception`_ v3
34
-  `GoogLeNet`_
Bar's avatar
Bar committed
35
-  `ShuffleNet`_ v2
36
37
-  `MobileNetV2`_
-  `MobileNetV3`_
38
-  `ResNeXt`_
39
-  `Wide ResNet`_
40
-  `MNASNet`_
41
-  `EfficientNet`_ v1 & v2
42
-  `RegNet`_
43
-  `VisionTransformer`_
44
-  `ConvNeXt`_
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
45
46
47
48
49
50
51
52
53
54

You can construct a model with random weights by calling its constructor:

.. code:: python

    import torchvision.models as models
    resnet18 = models.resnet18()
    alexnet = models.alexnet()
    vgg16 = models.vgg16()
    squeezenet = models.squeezenet1_0()
Ahmed Abdo's avatar
Ahmed Abdo committed
55
    densenet = models.densenet161()
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
56
    inception = models.inception_v3()
57
    googlenet = models.googlenet()
58
    shufflenet = models.shufflenet_v2_x1_0()
59
60
61
    mobilenet_v2 = models.mobilenet_v2()
    mobilenet_v3_large = models.mobilenet_v3_large()
    mobilenet_v3_small = models.mobilenet_v3_small()
62
    resnext50_32x4d = models.resnext50_32x4d()
63
    wide_resnet50_2 = models.wide_resnet50_2()
64
    mnasnet = models.mnasnet1_0()
65
66
67
68
69
70
71
72
    efficientnet_b0 = models.efficientnet_b0()
    efficientnet_b1 = models.efficientnet_b1()
    efficientnet_b2 = models.efficientnet_b2()
    efficientnet_b3 = models.efficientnet_b3()
    efficientnet_b4 = models.efficientnet_b4()
    efficientnet_b5 = models.efficientnet_b5()
    efficientnet_b6 = models.efficientnet_b6()
    efficientnet_b7 = models.efficientnet_b7()
73
74
75
    efficientnet_v2_s = models.efficientnet_v2_s()
    efficientnet_v2_m = models.efficientnet_v2_m()
    efficientnet_v2_l = models.efficientnet_v2_l()
76
77
78
79
80
81
82
    regnet_y_400mf = models.regnet_y_400mf()
    regnet_y_800mf = models.regnet_y_800mf()
    regnet_y_1_6gf = models.regnet_y_1_6gf()
    regnet_y_3_2gf = models.regnet_y_3_2gf()
    regnet_y_8gf = models.regnet_y_8gf()
    regnet_y_16gf = models.regnet_y_16gf()
    regnet_y_32gf = models.regnet_y_32gf()
83
    regnet_y_128gf = models.regnet_y_128gf()
84
85
86
87
88
89
90
    regnet_x_400mf = models.regnet_x_400mf()
    regnet_x_800mf = models.regnet_x_800mf()
    regnet_x_1_6gf = models.regnet_x_1_6gf()
    regnet_x_3_2gf = models.regnet_x_3_2gf()
    regnet_x_8gf = models.regnet_x_8gf()
    regnet_x_16gf = models.regnet_x_16gf()
    regnet_x_32gf = models.regnet_x_32gf()
91
92
93
94
    vit_b_16 = models.vit_b_16()
    vit_b_32 = models.vit_b_32()
    vit_l_16 = models.vit_l_16()
    vit_l_32 = models.vit_l_32()
95
96
97
98
    convnext_tiny = models.convnext_tiny()
    convnext_small = models.convnext_small()
    convnext_base = models.convnext_base()
    convnext_large = models.convnext_large()
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
99
100
101

We provide pre-trained models, using the PyTorch :mod:`torch.utils.model_zoo`.

102
Instancing a pre-trained model will download its weights to a cache directory.
103
104
This directory can be set using the `TORCH_HOME` environment variable. See
:func:`torch.hub.load_state_dict_from_url` for details.
105

106
107
108
Some models use modules which have different training and evaluation
behavior, such as batch normalization. To switch between these modes, use
``model.train()`` or ``model.eval()`` as appropriate. See
109
:meth:`~torch.nn.Module.train` or :meth:`~torch.nn.Module.eval` for details.
110

Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
111
112
113
114
115
116
117
118
119
120
121
122
123
All pre-trained models expect input images normalized in the same way,
i.e. mini-batches of 3-channel RGB images of shape (3 x H x W),
where H and W are expected to be at least 224.
The images have to be loaded in to a range of [0, 1] and then normalized
using ``mean = [0.485, 0.456, 0.406]`` and ``std = [0.229, 0.224, 0.225]``.
You can use the following transform to normalize::

    normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                     std=[0.229, 0.224, 0.225])

An example of such normalization can be found in the imagenet example
`here <https://github.com/pytorch/examples/blob/42e5b996718797e45c46a25c55b031e6768f8440/imagenet/main.py#L89-L101>`_

124
125
126
127
128
129
The process for obtaining the values of `mean` and `std` is roughly equivalent
to::

    import torch
    from torchvision import datasets, transforms as T

130
    transform = T.Compose([T.Resize(256), T.CenterCrop(224), T.PILToTensor(), T.ConvertImageDtype(torch.float)])
131
132
133
134
135
136
137
138
139
140
141
    dataset = datasets.ImageNet(".", split="train", transform=transform)

    means = []
    stds = []
    for img in subset(dataset):
        means.append(torch.mean(img))
        stds.append(torch.std(img))

    mean = torch.mean(torch.tensor(means))
    std = torch.mean(torch.tensor(stds))

142
Unfortunately, the concrete `subset` that was used is lost. For more
143
144
145
information see `this discussion <https://github.com/pytorch/vision/issues/1439>`_
or `these experiments <https://github.com/pytorch/vision/pull/1965>`_.

146
147
148
149
The sizes of the EfficientNet models depend on the variant. For the exact input sizes
`check here <https://github.com/pytorch/vision/blob/d2bfd639e46e1c5dc3c177f889dc7750c8d137c7/references/classification/train.py#L92-L93>`_

ImageNet 1-crop error rates
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
150
151

================================  =============   =============
152
Model                             Acc@1           Acc@5
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
153
================================  =============   =============
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
AlexNet                           56.522          79.066
VGG-11                            69.020          88.628
VGG-13                            69.928          89.246
VGG-16                            71.592          90.382
VGG-19                            72.376          90.876
VGG-11 with batch normalization   70.370          89.810
VGG-13 with batch normalization   71.586          90.374
VGG-16 with batch normalization   73.360          91.516
VGG-19 with batch normalization   74.218          91.842
ResNet-18                         69.758          89.078
ResNet-34                         73.314          91.420
ResNet-50                         76.130          92.862
ResNet-101                        77.374          93.546
ResNet-152                        78.312          94.046
SqueezeNet 1.0                    58.092          80.420
SqueezeNet 1.1                    58.178          80.624
Densenet-121                      74.434          91.972
Densenet-169                      75.600          92.806
Densenet-201                      76.896          93.370
Densenet-161                      77.138          93.560
Inception v3                      77.294          93.450
GoogleNet                         69.778          89.530
ShuffleNet V2 x1.0                69.362          88.316
ShuffleNet V2 x0.5                60.552          81.746
MobileNet V2                      71.878          90.286
MobileNet V3 Large                74.042          91.340
180
MobileNet V3 Small                67.668          87.402
181
182
183
184
185
186
ResNeXt-50-32x4d                  77.618          93.698
ResNeXt-101-32x8d                 79.312          94.526
Wide ResNet-50-2                  78.468          94.086
Wide ResNet-101-2                 78.848          94.284
MNASNet 1.0                       73.456          91.510
MNASNet 0.5                       67.734          87.490
187
188
189
190
191
192
193
194
EfficientNet-B0                   77.692          93.532
EfficientNet-B1                   78.642          94.186
EfficientNet-B2                   80.608          95.310
EfficientNet-B3                   82.008          96.054
EfficientNet-B4                   83.384          96.594
EfficientNet-B5                   83.444          96.628
EfficientNet-B6                   84.008          96.916
EfficientNet-B7                   84.122          96.908
195
196
197
EfficientNetV2-s                  84.228          96.878
EfficientNetV2-m                  85.112          97.156
EfficientNetV2-l                  85.810          97.792
198
199
200
201
202
203
204
205
regnet_x_400mf                    72.834          90.950
regnet_x_800mf                    75.212          92.348
regnet_x_1_6gf                    77.040          93.440
regnet_x_3_2gf                    78.364          93.992
regnet_x_8gf                      79.344          94.686 
regnet_x_16gf                     80.058          94.944
regnet_x_32gf                     80.622          95.248
regnet_y_400mf                    74.046          91.716
206
regnet_y_800mf                    76.420          93.136
207
208
209
210
211
regnet_y_1_6gf                    77.950          93.966
regnet_y_3_2gf                    78.948          94.576
regnet_y_8gf                      80.032          95.048
regnet_y_16gf                     80.424          95.240
regnet_y_32gf                     80.878          95.340
212
213
214
215
vit_b_16                          81.072          95.318
vit_b_32                          75.912          92.466
vit_l_16                          79.662          94.638
vit_l_32                          76.972          93.070
216
217
218
219
convnext_tiny                     82.520          96.146
convnext_small                    83.616          96.650
convnext_base                     84.062          96.870
convnext_large                    84.414          96.976
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
220
221
222
223
224
225
226
227
228
================================  =============   =============


.. _AlexNet: https://arxiv.org/abs/1404.5997
.. _VGG: https://arxiv.org/abs/1409.1556
.. _ResNet: https://arxiv.org/abs/1512.03385
.. _SqueezeNet: https://arxiv.org/abs/1602.07360
.. _DenseNet: https://arxiv.org/abs/1608.06993
.. _Inception: https://arxiv.org/abs/1512.00567
229
.. _GoogLeNet: https://arxiv.org/abs/1409.4842
Bar's avatar
Bar committed
230
.. _ShuffleNet: https://arxiv.org/abs/1807.11164
231
232
.. _MobileNetV2: https://arxiv.org/abs/1801.04381
.. _MobileNetV3: https://arxiv.org/abs/1905.02244
233
.. _ResNeXt: https://arxiv.org/abs/1611.05431
234
.. _MNASNet: https://arxiv.org/abs/1807.11626
235
.. _EfficientNet: https://arxiv.org/abs/1905.11946
236
.. _RegNet: https://arxiv.org/abs/2003.13678
237
.. _VisionTransformer: https://arxiv.org/abs/2010.11929
238
.. _ConvNeXt: https://arxiv.org/abs/2201.03545
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
239
240
241

.. currentmodule:: torchvision.models

Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
242
243
244
Alexnet
-------

245
246
247
248
249
.. autosummary::
    :toctree: generated/
    :template: function.rst

    alexnet
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
250
251
252
253

VGG
---

254
255
256
257
258
259
260
261
262
263
264
265
.. autosummary::
    :toctree: generated/
    :template: function.rst

    vgg11
    vgg11_bn
    vgg13
    vgg13_bn
    vgg16
    vgg16_bn
    vgg19
    vgg19_bn
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
266
267
268
269
270


ResNet
------

271
272
273
274
275
276
277
278
279
.. autosummary::
    :toctree: generated/
    :template: function.rst

    resnet18
    resnet34
    resnet50
    resnet101
    resnet152
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
280
281
282
283

SqueezeNet
----------

284
285
286
287
288
289
.. autosummary::
    :toctree: generated/
    :template: function.rst

    squeezenet1_0
    squeezenet1_1
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
290

Sangwhan Moon's avatar
Sangwhan Moon committed
291
DenseNet
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
292
293
---------

294
295
296
297
298
299
300
301
.. autosummary::
    :toctree: generated/
    :template: function.rst

    densenet121
    densenet169
    densenet161
    densenet201
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
302
303
304
305

Inception v3
------------

306
307
308
309
310
.. autosummary::
    :toctree: generated/
    :template: function.rst

    inception_v3
Sasank Chilamkurthy's avatar
Sasank Chilamkurthy committed
311

312
313
314
GoogLeNet
------------

315
316
317
318
319
.. autosummary::
    :toctree: generated/
    :template: function.rst

    googlenet
320

Bar's avatar
Bar committed
321
322
323
ShuffleNet v2
-------------

324
325
326
327
328
329
330
331
.. autosummary::
    :toctree: generated/
    :template: function.rst

    shufflenet_v2_x0_5
    shufflenet_v2_x1_0
    shufflenet_v2_x1_5
    shufflenet_v2_x2_0
Bar's avatar
Bar committed
332

333
334
335
MobileNet v2
-------------

336
337
338
339
340
.. autosummary::
    :toctree: generated/
    :template: function.rst

    mobilenet_v2
341

342
343
344
MobileNet v3
-------------

345
346
347
348
349
350
.. autosummary::
    :toctree: generated/
    :template: function.rst

    mobilenet_v3_large
    mobilenet_v3_small
351

352
ResNext
353
-------
354

355
356
357
358
359
360
.. autosummary::
    :toctree: generated/
    :template: function.rst

    resnext50_32x4d
    resnext101_32x8d
361

362
363
364
Wide ResNet
-----------

365
366
367
368
369
370
.. autosummary::
    :toctree: generated/
    :template: function.rst

    wide_resnet50_2
    wide_resnet101_2
371

372
373
374
MNASNet
--------

375
376
377
378
379
380
381
382
.. autosummary::
    :toctree: generated/
    :template: function.rst

    mnasnet0_5
    mnasnet0_75
    mnasnet1_0
    mnasnet1_3
383

384
385
386
EfficientNet
------------

387
388
389
390
391
392
393
394
395
396
397
398
.. autosummary::
    :toctree: generated/
    :template: function.rst

    efficientnet_b0
    efficientnet_b1
    efficientnet_b2
    efficientnet_b3
    efficientnet_b4
    efficientnet_b5
    efficientnet_b6
    efficientnet_b7
399
400
401
    efficientnet_v2_s
    efficientnet_v2_m
    efficientnet_v2_l
402

403
404
405
RegNet
------------

406
407
408
409
410
411
412
413
414
415
416
.. autosummary::
    :toctree: generated/
    :template: function.rst

    regnet_y_400mf
    regnet_y_800mf
    regnet_y_1_6gf
    regnet_y_3_2gf
    regnet_y_8gf
    regnet_y_16gf
    regnet_y_32gf
417
    regnet_y_128gf
418
419
420
421
422
423
424
    regnet_x_400mf
    regnet_x_800mf
    regnet_x_1_6gf
    regnet_x_3_2gf
    regnet_x_8gf
    regnet_x_16gf
    regnet_x_32gf
425

426
427
428
429
430
431
432
433
434
435
436
437
VisionTransformer
-----------------

.. autosummary::
    :toctree: generated/
    :template: function.rst

    vit_b_16
    vit_b_32
    vit_l_16
    vit_l_32

438
439
440
441
442
443
444
445
446
447
448
449
ConvNeXt
--------

.. autosummary::
    :toctree: generated/
    :template: function.rst

    convnext_tiny
    convnext_small
    convnext_base
    convnext_large

450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
Quantized Models
----------------

The following architectures provide support for INT8 quantized models. You can get
a model with random weights by calling its constructor:

.. code:: python

    import torchvision.models as models
    googlenet = models.quantization.googlenet()
    inception_v3 = models.quantization.inception_v3()
    mobilenet_v2 = models.quantization.mobilenet_v2()
    mobilenet_v3_large = models.quantization.mobilenet_v3_large()
    resnet18 = models.quantization.resnet18()
    resnet50 = models.quantization.resnet50()
    resnext101_32x8d = models.quantization.resnext101_32x8d()
    shufflenet_v2_x0_5 = models.quantization.shufflenet_v2_x0_5()
    shufflenet_v2_x1_0 = models.quantization.shufflenet_v2_x1_0()

Obtaining a pre-trained quantized model can be done with a few lines of code:

.. code:: python

    import torchvision.models as models
474
    model = models.quantization.mobilenet_v2(weights=MobileNet_V2_QuantizedWeights.IMAGENET1K_QNNPACK_V1, quantize=True)
475
476
477
478
479
480
481
482
483
484
485
    model.eval()
    # run the model with quantized inputs and weights
    out = model(torch.rand(1, 3, 224, 224))

We provide pre-trained quantized weights for the following models:

================================  =============  =============
Model                             Acc@1          Acc@5
================================  =============  =============
MobileNet V2                      71.658         90.150
MobileNet V3 Large                73.004         90.858
486
487
ShuffleNet V2 x1.0                68.360         87.582
ShuffleNet V2 x0.5                57.972         79.780
488
489
490
491
492
493
494
ResNet 18                         69.494         88.882
ResNet 50                         75.920         92.814
ResNext 101 32x8d                 78.986         94.480
Inception V3                      77.176         93.354
GoogleNet                         69.826         89.404
================================  =============  =============

495
496
497
498

Semantic Segmentation
=====================

499
500
501
The models subpackage contains definitions for the following model
architectures for semantic segmentation:

502
- `FCN ResNet50, ResNet101 <https://arxiv.org/abs/1411.4038>`_
503
504
- `DeepLabV3 ResNet50, ResNet101, MobileNetV3-Large <https://arxiv.org/abs/1706.05587>`_
- `LR-ASPP MobileNetV3-Large <https://arxiv.org/abs/1905.02244>`_
505

506
507
508
509
510
As with image classification models, all pre-trained models expect input images normalized in the same way.
The images have to be loaded in to a range of ``[0, 1]`` and then normalized using
``mean = [0.485, 0.456, 0.406]`` and ``std = [0.229, 0.224, 0.225]``.
They have been trained on images resized such that their minimum size is 520.

511
512
For details on how to plot the masks of such models, you may refer to :ref:`semantic_seg_output`.

513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
The pre-trained models have been trained on a subset of COCO train2017, on the 20 categories that are
present in the Pascal VOC dataset. You can see more information on how the subset has been selected in
``references/segmentation/coco_utils.py``. The classes that the pre-trained model outputs are the following,
in order:

  .. code-block:: python

      ['__background__', 'aeroplane', 'bicycle', 'bird', 'boat', 'bottle', 'bus',
       'car', 'cat', 'chair', 'cow', 'diningtable', 'dog', 'horse', 'motorbike',
       'person', 'pottedplant', 'sheep', 'sofa', 'train', 'tvmonitor']

The accuracies of the pre-trained models evaluated on COCO val2017 are as follows

================================  =============  ====================
Network                           mean IoU       global pixelwise acc
================================  =============  ====================
529
FCN ResNet50                      60.5           91.4
530
FCN ResNet101                     63.7           91.9
531
DeepLabV3 ResNet50                66.4           92.4
532
DeepLabV3 ResNet101               67.4           92.4
533
534
DeepLabV3 MobileNetV3-Large       60.3           91.2
LR-ASPP MobileNetV3-Large         57.9           91.2
535
536
537
538
539
540
================================  =============  ====================


Fully Convolutional Networks
----------------------------

541
542
543
544
545
546
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.segmentation.fcn_resnet50
    torchvision.models.segmentation.fcn_resnet101
547
548
549
550
551


DeepLabV3
---------

552
553
554
555
556
557
558
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.segmentation.deeplabv3_resnet50
    torchvision.models.segmentation.deeplabv3_resnet101
    torchvision.models.segmentation.deeplabv3_mobilenet_v3_large
559
560
561
562
563


LR-ASPP
-------

564
565
566
567
568
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.segmentation.lraspp_mobilenet_v3_large
569

570
.. _object_det_inst_seg_pers_keypoint_det:
571
572
573
574

Object Detection, Instance Segmentation and Person Keypoint Detection
=====================================================================

575
576
577
The models subpackage contains definitions for the following model
architectures for detection:

578
- `Faster R-CNN <https://arxiv.org/abs/1506.01497>`_
Hu Ye's avatar
Hu Ye committed
579
- `FCOS <https://arxiv.org/abs/1904.01355>`_
580
581
582
- `Mask R-CNN <https://arxiv.org/abs/1703.06870>`_
- `RetinaNet <https://arxiv.org/abs/1708.02002>`_
- `SSD <https://arxiv.org/abs/1512.02325>`_
583
- `SSDlite <https://arxiv.org/abs/1801.04381>`_
584

585
586
587
588
589
The pre-trained models for detection, instance segmentation and
keypoint detection are initialized with the classification models
in torchvision.

The models expect a list of ``Tensor[C, H, W]``, in the range ``0-1``.
590
The models internally resize the images but the behaviour varies depending
591
592
on the model. Check the constructor of the models for more information. The
output format of such models is illustrated in :ref:`instance_seg_output`.
593
594
595
596
597
598
599
600
601


For object detection and instance segmentation, the pre-trained
models return the predictions of the following classes:

  .. code-block:: python

      COCO_INSTANCE_CATEGORY_NAMES = [
          '__background__', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus',
602
603
          'train', 'truck', 'boat', 'traffic light', 'fire hydrant', 'N/A', 'stop sign',
          'parking meter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow',
604
          'elephant', 'bear', 'zebra', 'giraffe', 'N/A', 'backpack', 'umbrella', 'N/A', 'N/A',
605
606
607
608
609
610
611
612
          'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball',
          'kite', 'baseball bat', 'baseball glove', 'skateboard', 'surfboard', 'tennis racket',
          'bottle', 'N/A', 'wine glass', 'cup', 'fork', 'knife', 'spoon', 'bowl',
          'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hot dog', 'pizza',
          'donut', 'cake', 'chair', 'couch', 'potted plant', 'bed', 'N/A', 'dining table',
          'N/A', 'N/A', 'toilet', 'N/A', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cell phone',
          'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'N/A', 'book',
          'clock', 'vase', 'scissors', 'teddy bear', 'hair drier', 'toothbrush'
613
614
615
616
617
618
      ]


Here are the summary of the accuracies for the models trained on
the instances set of COCO train2017 and evaluated on COCO val2017.

619
620
621
622
623
624
======================================  =======  ========  ===========
Network                                 box AP   mask AP   keypoint AP
======================================  =======  ========  ===========
Faster R-CNN ResNet-50 FPN              37.0     -         -
Faster R-CNN MobileNetV3-Large FPN      32.8     -         -
Faster R-CNN MobileNetV3-Large 320 FPN  22.8     -         -
Hu Ye's avatar
Hu Ye committed
625
FCOS ResNet-50 FPN                      39.2     -         -
626
RetinaNet ResNet-50 FPN                 36.4     -         -
627
628
SSD300 VGG16                            25.1     -         -
SSDlite320 MobileNetV3-Large            21.3     -         -
629
630
Mask R-CNN ResNet-50 FPN                37.9     34.6      -
======================================  =======  ========  ===========
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665

For person keypoint detection, the accuracies for the pre-trained
models are as follows

================================  =======  ========  ===========
Network                           box AP   mask AP   keypoint AP
================================  =======  ========  ===========
Keypoint R-CNN ResNet-50 FPN      54.6     -         65.0
================================  =======  ========  ===========

For person keypoint detection, the pre-trained model return the
keypoints in the following order:

  .. code-block:: python

    COCO_PERSON_KEYPOINT_NAMES = [
        'nose',
        'left_eye',
        'right_eye',
        'left_ear',
        'right_ear',
        'left_shoulder',
        'right_shoulder',
        'left_elbow',
        'right_elbow',
        'left_wrist',
        'right_wrist',
        'left_hip',
        'right_hip',
        'left_knee',
        'right_knee',
        'left_ankle',
        'right_ankle'
    ]

666
667
668
669
670
671
Runtime characteristics
-----------------------

The implementations of the models for object detection, instance segmentation
and keypoint detection are efficient.

672
673
674
In the following table, we use 8 GPUs to report the results. During training,
we use a batch size of 2 per GPU for all models except SSD which uses 4
and SSDlite which uses 24. During testing a batch size  of 1 is used.
675
676
677
678
679

For test time, we report the time for the model evaluation and postprocessing
(including mask pasting in image), but not the time for computing the
precision-recall.

680
681
682
683
684
685
======================================  ===================  ==================  ===========
Network                                 train time (s / it)  test time (s / it)  memory (GB)
======================================  ===================  ==================  ===========
Faster R-CNN ResNet-50 FPN              0.2288               0.0590              5.2
Faster R-CNN MobileNetV3-Large FPN      0.1020               0.0415              1.0
Faster R-CNN MobileNetV3-Large 320 FPN  0.0978               0.0376              0.6
Hu Ye's avatar
Hu Ye committed
686
FCOS ResNet-50 FPN                      0.1450               0.0539              3.3
687
RetinaNet ResNet-50 FPN                 0.2514               0.0939              4.1
688
689
SSD300 VGG16                            0.2093               0.0744              1.5
SSDlite320 MobileNetV3-Large            0.1773               0.0906              1.5
690
691
692
Mask R-CNN ResNet-50 FPN                0.2728               0.0903              5.4
Keypoint R-CNN ResNet-50 FPN            0.3789               0.1242              6.8
======================================  ===================  ==================  ===========
693
694
695
696
697


Faster R-CNN
------------

698
699
700
701
702
703
704
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.detection.fasterrcnn_resnet50_fpn
    torchvision.models.detection.fasterrcnn_mobilenet_v3_large_fpn
    torchvision.models.detection.fasterrcnn_mobilenet_v3_large_320_fpn
705

Hu Ye's avatar
Hu Ye committed
706
707
708
709
710
711
712
713
714
FCOS
----

.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.detection.fcos_resnet50_fpn

715

716
RetinaNet
717
---------
718

719
720
721
722
723
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.detection.retinanet_resnet50_fpn
724
725


726
SSD
727
---
728

729
730
731
732
733
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.detection.ssd300_vgg16
734
735


736
SSDlite
737
-------
738

739
740
741
742
743
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.detection.ssdlite320_mobilenet_v3_large
744
745


746
747
748
Mask R-CNN
----------

749
750
751
752
753
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.detection.maskrcnn_resnet50_fpn
754
755
756
757
758


Keypoint R-CNN
--------------

759
760
761
762
763
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.detection.keypointrcnn_resnet50_fpn
764

765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801

Video classification
====================

We provide models for action recognition pre-trained on Kinetics-400.
They have all been trained with the scripts provided in ``references/video_classification``.

All pre-trained models expect input images normalized in the same way,
i.e. mini-batches of 3-channel RGB videos of shape (3 x T x H x W),
where H and W are expected to be 112, and T is a number of video frames in a clip.
The images have to be loaded in to a range of [0, 1] and then normalized
using ``mean = [0.43216, 0.394666, 0.37645]`` and ``std = [0.22803, 0.22145, 0.216989]``.


.. note::
  The normalization parameters are different from the image classification ones, and correspond
  to the mean and std from Kinetics-400.

.. note::
  For now, normalization code can be found in ``references/video_classification/transforms.py``,
  see the ``Normalize`` function there. Note that it differs from standard normalization for
  images because it assumes the video is 4d.

Kinetics 1-crop accuracies for clip length 16 (16x112x112)

================================  =============   =============
Network                           Clip acc@1      Clip acc@5
================================  =============   =============
ResNet 3D 18                      52.75           75.45
ResNet MC 18                      53.90           76.29
ResNet (2+1)D                     57.50           78.81
================================  =============   =============


ResNet 3D
----------

802
803
804
805
806
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.video.r3d_18
807
808
809
810

ResNet Mixed Convolution
------------------------

811
812
813
814
815
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.video.mc3_18
816
817
818
819

ResNet (2+1)D
-------------

820
821
822
823
824
.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.video.r2plus1d_18
825
826
827
828
829
830
831
832
833
834
835
836
837

Optical flow
============

Raft
----

.. autosummary::
    :toctree: generated/
    :template: function.rst

    torchvision.models.optical_flow.raft_large
    torchvision.models.optical_flow.raft_small