Replace top-X error with Acc@X and fix metric discrepancies. (#3360)

17e1ec4e · Vasilis Vryniotis · GitHub · ced96a0c · 17e1ec4e
Unverified Commit 17e1ec4e authored Feb 08, 2021 by Vasilis Vryniotis Committed by GitHub Feb 08, 2021
Hide whitespace changes
Inline Side-by-side

Showing with 33 additions and 31 deletions

docs/source/models.rst docs/source/models.rst +33 -31

No files found.
--- a/docs/source/models.rst
+++ b/docs/source/models.rst
@@ -115,38 +115,40 @@ or `these experiments <https://github.com/pytorch/vision/pull/1965>`_.
 ImageNet 1-crop error rates (224x224)
 ================================  =============   =============
-Network                           Top-1 error     Top-5 error
+Model                             Acc@1           Acc@5
 ================================  =============   =============
-AlexNet                           43.45           20.91
+AlexNet                           56.522          79.066
-VGG-11                            30.98           11.37
+VGG-11                            69.020          88.628
-VGG-13                            30.07           10.75
+VGG-13                            69.928          89.246
-VGG-16                            28.41           9.62
+VGG-16                            71.592          90.382
-VGG-19                            27.62           9.12
+VGG-19                            72.376          90.876
-VGG-11 with batch normalization   29.62           10.19
+VGG-11 with batch normalization   70.370          89.810
-VGG-13 with batch normalization   28.45           9.63
+VGG-13 with batch normalization   71.586          90.374
-VGG-16 with batch normalization   26.63           8.50
+VGG-16 with batch normalization   73.360          91.516
-VGG-19 with batch normalization   25.76           8.15
+VGG-19 with batch normalization   74.218          91.842
-ResNet-18                         30.24           10.92
+ResNet-18                         69.758          89.078
-ResNet-34                         26.70           8.58
+ResNet-34                         73.314          91.420
-ResNet-50                         23.85           7.13
+ResNet-50                         76.130          92.862
-ResNet-101                        22.63           6.44
+ResNet-101                        77.374          93.546
-ResNet-152                        21.69           5.94
+ResNet-152                        78.312          94.046
-SqueezeNet 1.0                    41.90           19.58
+SqueezeNet 1.0                    58.092          80.420
-SqueezeNet 1.1                    41.81           19.38
+SqueezeNet 1.1                    58.178          80.624
-Densenet-121                      25.35           7.83
+Densenet-121                      74.434          91.972
-Densenet-169                      24.00           7.00
+Densenet-169                      75.600          92.806
-Densenet-201                      22.80           6.43
+Densenet-201                      76.896          93.370
-Densenet-161                      22.35           6.20
+Densenet-161                      77.138          93.560
-Inception v3                      22.55           6.44
+Inception v3                      77.294          93.450
-GoogleNet                         30.22           10.47
+GoogleNet                         69.778          89.530
-ShuffleNet V2                     30.64           11.68
+ShuffleNet V2 x1.0                69.362          88.316
-MobileNet V2                      28.12           9.71
+ShuffleNet V2 x0.5                60.552          81.746
-MobileNet V3 Large                25.96           8.66
+MobileNet V2                      71.878          90.286
-ResNeXt-50-32x4d                  22.38           6.30
+MobileNet V3 Large                74.042          91.340
-ResNeXt-101-32x8d                 20.69           5.47
+ResNeXt-50-32x4d                  77.618          93.698
-Wide ResNet-50-2                  21.49           5.91
+ResNeXt-101-32x8d                 79.312          94.526
-Wide ResNet-101-2                 21.16           5.72
+Wide ResNet-50-2                  78.468          94.086
-MNASNet 1.0                       26.49           8.456
+Wide ResNet-101-2                 78.848          94.284
+MNASNet 1.0                       73.456          91.510
+MNASNet 0.5                       67.734          87.490
 ================================  =============   =============