1. Home
  2. rwightman
  3. rwightman/pytorch-image-models
  4. Build #24

rwightman / pytorch-image-models » Build #24 

Paper not specified for task=`Image Classification` on dataset=`ImageNet`
Paper not specified for task=`Image Classification` on dataset=`ImageNet`

Build Results

Dataset Image Classification on ImageNet

Image Classification on ImageNet

TOP 1 ACCURACY TOP 5 ACCURACY
SPEED
MODEL CODE PAPER
ε-REPR
CODE PAPER
ε-REPR
PAPER
Adversarial Inception V3
77.6% -- 93.7% -- 269.1
DLA-102
78.0% -- 94.0% -- 384.0
DLA-169
78.7% -- 94.3% -- 368.5
DLA-34
74.6% -- 92.1% -- 404.7
DLA-46-C
64.9% -- 86.3% -- 407.5
DLA-60
77.0% -- 93.3% -- 391.0
DLA-X-102
78.5% -- 94.2% -- 381.6
DLA-X-102 64
79.5% -- 94.6% -- 348.3
DLA-X-46-C
66.0% -- 87.0% -- 405.5
DLA-X-60
78.2% -- 94.0% -- 387.5
DLA-X-60-C
67.9% -- 88.4% -- 404.8
DPN-107
(224x224)
80.2% -- 94.9% -- 215.4
DPN-107
(320x320, Mean-Max Pooling)
81.8% -- 95.9% -- 106.6
DPN-131
(224x224)
79.8% 80.1% 94.7% 94.9% 240.0
DPN-131
(320x320, Mean-Max Pooling)
81.4% 81.5% 95.8% 95.8% 118.6
DPN-68
(224x224)
76.3% 76.4% 93.0% 93.1% 357.0
DPN-68
(320x320, Mean-Max Pooling)
78.5% 78.5% 94.4% 94.5% 292.3
DPN-68b
(224x224)
77.5% -- 93.8% -- 363.5
DPN-68b
(320x320, Mean-Max Pooling)
79.4% -- 95.0% -- 295.3
DPN-92
(224x224)
80.0% 79.3%
94.8% 94.6% 338.0
DPN-92
(320x320, Mean-Max Pooling)
81.3% 81.0%
95.7% 95.5% 244.4
DPN-98
(224x224)
79.6% 80.0% 94.6% 94.8% 317.9
DPN-98
(320x320, Mean-Max Pooling)
81.2% 81.3% 95.7% 95.6% 164.0
EfficientNet-B0
76.9% 76.3%
93.2% 93.2% 363.4
EfficientNet-B0
(AdvProp)
77.1% -- 93.3% -- 366.9
EfficientNet-B0
(AutoAugment)
76.8% -- 93.2% -- 358.6
EfficientNet-B1
78.7% 78.8% 94.2% 94.4% 309.3
EfficientNet-B1
(AdvProp)
79.3% -- 94.3% -- 349.3
EfficientNet-B1
(AutoAugment)
78.8% -- 94.2% -- 339.9
EfficientNet-B2
80.4% 79.8%
95.1% 94.9% 336.5
EfficientNet-B2
(288x288, 1.0 crop)
80.6% -- 95.3% -- 325.1
EfficientNet-B2
(AdvProp)
80.3% -- 95.0% -- 333.1
EfficientNet-B2
(AutoAugment)
80.1% -- 94.9% -- 325.8
EfficientNet-B3
81.5% 81.1%
95.7% 95.5% 304.0
EfficientNet-B3
(320x320, 1.0 crop)
81.9% -- 95.8% -- 293.9
EfficientNet-B3
(AdvProp)
81.8% -- 95.6% -- 303.7
EfficientNet-B3
(AutoAugment)
81.6% -- 95.7% -- 300.5
EfficientNet-B4
(AdvProp)
83.2% -- 96.4% -- 240.2
EfficientNet-B4
(AutoAugment)
83.0% -- 96.3% -- 234.5
EfficientNet-B5
(AdvProp)
84.3% -- 97.0% -- 146.8
EfficientNet-B5
(RandAugment)
83.8% -- 96.8% -- 126.4
EfficientNet-B6
(AdvProp)
84.8% -- 97.1% -- 86.1
EfficientNet-B6
(AutoAugment)
84.1% -- 96.9% -- 73.6
EfficientNet-B7
(AdvProp)
85.1% -- 97.3% -- 50.1
EfficientNet-B7
(RandAugment)
84.9% -- 97.2% -- 43.1
EfficientNet-B8
(AdvProp)
85.4% -- 97.3% -- 32.8
EfficientNet-CondConv-B0 4 experts
77.3% -- 93.3% -- 364.5
EfficientNet-CondConv-B0 8 experts
77.9% -- 93.7% -- 360.9
EfficientNet-CondConv-B1 8 experts
79.3% -- 94.4% -- 348.2
EfficientNet-EdgeTPU-L
80.4% -- 95.2% -- 282.0
EfficientNet-EdgeTPU-M
78.7% -- 94.3% -- 336.5
EfficientNet-EdgeTPU-S
77.3% -- 93.6% -- 358.9
Ensemble Adversarial Inception V3
80.0% -- 94.9% -- 289.2
FBNet-C
75.1% 74.9% 92.4% -- 422.1
HRNet-W18-C
76.8% -- 93.4% -- 389.9
HRNet-W18-C-Small-V1
72.3% -- 90.7% -- 409.3
HRNet-W18-C-Small-V2
75.1% -- 92.4% -- 402.9
HRNet-W30-C
78.2% -- 94.2% -- 387.4
HRNet-W32-C
78.4% -- 94.2% -- 378.9
HRNet-W40-C
78.9% -- 94.5% -- 355.8
HRNet-W44-C
78.9% -- 94.4% -- 336.9
HRNet-W48-C
79.3% -- 94.5% -- 315.7
HRNet-W64-C
79.5% -- 94.7% -- 265.4
Inception ResNet V2
80.5% 80.1%
95.3% 95.1% 287.1
Inception V3
78.8% 78.8% 93.6% 94.4% 303.2
Inception V4
80.2% -- 95.0% -- 288.0
MixNet-L
79.0% 78.9% 94.0% 94.2% 356.0
MixNet-M
77.0% 77.0% 93.4% 93.3% 357.4
MixNet-S
75.6% 75.8% 92.8% 92.8% 360.3
MixNet-XL
80.5% -- 94.9% -- 355.3
MnasNet-A1
75.5% 75.2% 92.6% 92.5% 356.2
MnasNet-B1
74.7% -- 92.1% -- 356.9
MobileNet V3-Large 0.75
73.4% -- 91.4% -- 410.5
MobileNet V3-Large 1.0
75.5% 75.2%
92.6% -- 417.6
MobileNet V3-Large Minimal 1.0
72.2% -- 90.6% -- 410.5
MobileNet V3-Small 0.75
65.7% -- 86.1% -- 422.9
MobileNet V3-Small 1.0
67.9% -- 87.7% -- 414.6
MobileNet V3-Small Minimal 1.0
62.9% -- 84.2% -- 414.4
Modified Aligned Xception
79.6% 79.8% 94.7% 94.8% 260.0
NASNet-A Large
82.6% -- 96.0% -- 116.7
PNASNet-5
82.7% 82.9% 96.0% 96.2% 118.1
Res2Net-50 14x8s
78.2% -- 93.8% -- 386.0
Res2Net-50 26x4s
77.9% -- 93.9% -- 394.0
Res2Net-50 26x6s
78.6% -- 94.1% -- 382.9
Res2Net-50 26x8s
79.2% -- 94.4% -- 377.1
Res2Net-50 48x2s
77.5% -- 93.5% -- 386.9
Res2Net-DLA-60
78.5% 79.5% 94.2% -- 391.6
Res2NeXt-101 26x4s
79.2% -- 94.4% -- 377.8
Res2NeXt-50
78.2% -- 93.9% -- 377.0
Res2NeXt-DLA-60
78.4% -- 94.1% -- 377.1
ResNet-101
79.3% -- 94.5% -- 346.2
ResNet-101-C
79.5% -- 94.6% -- 349.5
ResNet-101-D
80.4% -- 95.0% -- 346.8
ResNet-101-S
80.3% -- 95.2% -- 341.9
ResNet-152
79.7% -- 94.7% -- 339.9
ResNet-152-C
79.9% -- 94.8% -- 333.1
ResNet-152-D
80.5% -- 95.2% -- 337.3
ResNet-152-S
81.0% -- 95.4% -- 335.2
ResNet-18
73.3% -- 91.7% -- 413.6
ResNet-18
70.8% -- 89.8% -- 368.7
ResNet-26
75.3% -- 92.6% -- 354.9
ResNet-26-D
76.7% -- 93.2% -- 352.7
ResNet-34
75.1% -- 92.3% -- 416.9
ResNet-50
77.6% -- 94.3% -- 355.0
ResNet-50
81.2% -- 96.0% -- 397.4
ResNet-50
(288x288 Mean-Max Pooling)
80.3% -- 95.6% -- 346.0
ResNet-50-C
78.0% -- 94.0% -- 362.6
ResNet-50-D
79.1% 77.2%
94.5% 93.5%
357.3
ResNet-50-S
78.7% -- 94.2% -- 352.4
ResNeXt-101 32x16d
81.8% -- 96.1% -- 150.9
ResNeXt-101 32x16d
84.2% -- 97.2% -- 150.9
ResNeXt-101 32x16d
(288x288 Mean-Max Pooling)
84.3% -- 97.3% -- 81.6
ResNeXt-101 32x16d
(288x288 Mean-Max Pooling)
85.0% -- 97.6% -- 84.1
ResNeXt-101 32x32d
85.1% 85.1% 97.4% 97.5% 60.5
ResNeXt-101 32x32d
(288x288 Mean-Max Pooling)
85.9% -- 97.8% -- 37.6
ResNeXt-101 32x48d
85.4% 85.4% 97.6% 97.6% 31.6
ResNeXt-101 32x48d
(288x288 Mean-Max Pooling)
86.1% -- 97.9% -- 19.1
ResNeXt-101 32x4d
80.9% -- 95.7% -- 372.3
ResNeXt-101 32x4d
80.3% -- 94.9% -- 342.0
ResNeXt-101 32x4d
(288x288 Mean-Max Pooling)
82.0% -- 96.3% -- 283.3
ResNeXt-101 32x8d
84.3% -- 97.2% -- 286.4
ResNeXt-101 32x8d
82.7% 82.2%
96.6% 96.4% 286.2
ResNeXt-101 32x8d
(288x288 Mean-Max Pooling)
85.1% -- 96.5% -- 175.3
ResNeXt-101 32x8d
(288x288 Mean-Max Pooling)
83.5% -- 97.1% -- 174.6
ResNeXt-101 64x4d
80.6% -- 95.0% -- 277.5
ResNeXt-50 32x4d
79.4% -- 94.1% -- 345.2
ResNeXt-50 32x4d
80.3% -- 96.2% -- 393.7
ResNeXt-50 32x4d
(288x288 Mean-Max Pooling)
83.1% -- 96.8% -- 337.8
ResNeXt-50-D 32x4d
79.7% -- 94.9% -- 348.5
SENet-154
81.2% -- 95.4% -- 214.8
SENet-154
81.3% 82.7% 95.5% 96.2% 215.6
SE-ResNet-101
78.4% -- 94.3% -- 391.5
SE-ResNet-152
78.7% -- 94.4% -- 371.8
SE-ResNet-18
71.8% -- 90.3% -- 354.7
SE-ResNet-34
74.8% -- 92.1% -- 410.7
SE-ResNet-50
77.6% -- 93.8% -- 399.1
SE-ResNeXt-101 32x4d
80.2% -- 95.0% -- 375.5
SE-ResNeXt-101 32x4d
80.9% -- 95.3% -- 335.2
SE-ResNeXt-101 64x4d
80.9% -- 95.3% -- 266.0
SE-ResNeXt-26 32x4d
77.1% -- 93.3% -- 353.9
SE-ResNeXt-50 32x4d
79.1% -- 94.4% -- 393.5
SE-ResNeXt-50 32x4d
79.9% -- 94.8% -- 347.1
Single-Path NAS
74.1% 75.0% 91.8% 92.2% 417.0
Xception
79.0% 79.0% 94.4% 94.5% 295.6