rwightman / pytorch-image-models

TOP 1 ACCURACY TOP 5 ACCURACY
SPEED
MODEL CODE PAPER
ε-REPR
CODE PAPER
ε-REPR
PAPER
GLOBAL RANK
Adversarial Inception V3
77.6% -- 93.7% -- 269.1 #300
DenseNet-121
75.6% 76.4% 92.7% 93.3% 366.6 #372
DenseNet-Blur-121D
76.6% -- 93.2% -- 365.9 #354
DLA-102
78.0% -- 94.0% -- 384.0 #288
DLA-169
78.7% -- 94.3% -- 368.5 #239
DLA-34
74.6% -- 92.1% -- 404.7 #400
DLA-46-C
64.9% -- 86.3% -- 407.5 #529
DLA-60
77.0% -- 93.3% -- 391.0 #335
DLA-X-102
78.5% -- 94.2% -- 381.6 #245
DLA-X-102 64
79.5% -- 94.6% -- 348.3 #184
DLA-X-46-C
66.0% -- 87.0% -- 405.5 #518
DLA-X-60
78.2% -- 94.0% -- 387.5 #272
DLA-X-60-C
67.9% -- 88.4% -- 404.8 #495
DPN-107
(224x224)
80.2% -- 94.9% -- 215.4 #149
DPN-107
(320x320, Mean-Max Pooling)
81.8% -- 95.9% -- 106.6 #90
DPN-131
(224x224)
79.8% 80.1% 94.7% 94.9% 240.0 #179
DPN-131
(320x320, Mean-Max Pooling)
81.4% 81.5% 95.8% 95.8% 118.6 #98
DPN-68
(224x224)
76.3% 76.4% 93.0% 93.1% 357.0 #361
DPN-68
(320x320, Mean-Max Pooling)
78.5% 78.5% 94.4% 94.5% 292.3 #225
DPN-68b
(224x224)
77.5% -- 93.8% -- 363.5 #294
DPN-68b
(320x320, Mean-Max Pooling)
79.4% -- 95.0% -- 295.3 #186
DPN-92
(224x224)
80.0% 79.3%
94.8% 94.6% 338.0 #159
DPN-92
(320x320, Mean-Max Pooling)
81.3% 81.0%
95.7% 95.5% 244.4 #102
DPN-98
(224x224)
79.6% 80.0% 94.6% 94.8% 317.9 #192
DPN-98
(320x320, Mean-Max Pooling)
81.2% 81.3% 95.7% 95.6% 164.0 #104
ECA-ResNet-101d
82.2% -- 96.1% -- #82
ECA-ResNet-50d
80.6% -- 95.3% -- #126
ECA-ResNet-Light
80.5% -- 95.3% -- #131
EfficientNet-B0
77.7% 76.3%
93.5% 93.2%
357.6 #302
EfficientNet-B0
(AdvProp)
77.1% -- 93.3% -- 366.9 #346
EfficientNet-B0
(AutoAugment)
76.8% -- 93.2% -- 358.6 #348
EfficientNet-B0
(NoisyStudent)
78.7% -- 94.4% -- 359.1 #229
EfficientNet-B1
78.7% 78.8% 94.2% 94.4% 309.3 #241
EfficientNet-B1
(AdvProp)
79.3% -- 94.3% -- 349.3 #198
EfficientNet-B1
(AutoAugment)
78.8% -- 94.2% -- 339.9 #252
EfficientNet-B1
(NoisyStudent)
81.4% -- 95.7% -- 350.6 #97
EfficientNet-B2
80.4% 79.8%
95.1% 94.9% 336.5 #141
EfficientNet-B2
(288x288, 1.0 crop)
80.6% -- 95.3% -- 325.1 #125
EfficientNet-B2
(AdvProp)
80.3% -- 95.0% -- 333.1 #143
EfficientNet-B2
(AutoAugment)
80.1% -- 94.9% -- 325.8 #159
EfficientNet-B2
(NoisyStudent)
82.4% -- 96.3% -- 335.6 #78
EfficientNet-B3
81.5% 81.1%
95.7% 95.5% 304.0 #97
EfficientNet-B3
(320x320, 1.0 crop)
81.9% -- 95.8% -- 293.9 #88
EfficientNet-B3
(AdvProp)
81.8% -- 95.6% -- 303.7 #91
EfficientNet-B3
(AutoAugment)
81.6% -- 95.7% -- 300.5 #93
EfficientNet-B3
(NoisyStudent)
84.1% -- 96.9% -- 306.6 #47
EfficientNet-B4
(AdvProp)
83.2% -- 96.4% -- 240.2 #70
EfficientNet-B4
(AutoAugment)
83.0% -- 96.3% -- 234.5 #74
EfficientNet-B4
(NoisyStudent)
85.2% -- 97.5% -- 241.1 #22
EfficientNet-B5
(AdvProp)
84.3% -- 97.0% -- 146.8 #43
EfficientNet-B5
(NoisyStudent)
86.1% -- 97.8% -- 147.5 #11
EfficientNet-B5
(RandAugment)
83.8% -- 96.8% -- 126.4 #53
EfficientNet-B6
(AdvProp)
84.8% -- 97.1% -- 86.1 #39
EfficientNet-B6
(AutoAugment)
84.1% -- 96.9% -- 73.6 #44
EfficientNet-B6
(NoisyStudent)
86.5% -- 97.9% -- 86.7 #10
EfficientNet-B7
(AdvProp)
85.1% -- 97.3% -- 50.1 #24
EfficientNet-B7
(NoisyStudent)
86.8% -- 98.1% -- 50.4 #5
EfficientNet-B7
(RandAugment)
84.9% -- 97.2% -- 43.1 #34
EfficientNet-B8
(AdvProp)
85.4% -- 97.3% -- 32.8 #18
EfficientNet-B8
(RandAugment)
85.4% -- 97.4% -- 33.0 #27
EfficientNet-CondConv-B0 4 experts
77.3% -- 93.3% -- 364.5 #320
EfficientNet-CondConv-B0 8 experts
77.9% -- 93.7% -- 360.9 #296
EfficientNet-CondConv-B1 8 experts
79.3% -- 94.4% -- 348.2 #230
EfficientNet-EdgeTPU-L
80.4% -- 95.2% -- 282.0 #132
EfficientNet-EdgeTPU-M
78.7% -- 94.3% -- 336.5 #236
EfficientNet-EdgeTPU-S
77.3% -- 93.6% -- 358.9 #313
EfficientNet-L2 475
(NoisyStudent)
88.2% -- 98.5% -- 17.4 #3
EfficientNet-L2
(NoisyStudent)
88.4% -- 98.6% -- 6.5 #2
EfficientNet-Lite0
74.8% -- 92.2% -- 356.6 #396
EfficientNet-Lite1
76.6% -- 93.2% -- 340.7 #347
EfficientNet-Lite2
77.5% -- 93.7% -- 321.7 #316
EfficientNet-Lite3
79.8% -- 94.9% -- 333.5 #156
EfficientNet-Lite4
81.5% -- 95.7% -- 264.8 #103
Ensemble Adversarial Inception V3
80.0% -- 94.9% -- 289.2 #152
FBNet-C
75.1% 74.9% 92.4% -- 422.1 #386
HRNet-W18-C
76.8% -- 93.4% -- 389.9 #327
HRNet-W18-C-Small-V1
72.3% -- 90.7% -- 409.3 #444
HRNet-W18-C-Small-V2
75.1% -- 92.4% -- 402.9 #383
HRNet-W30-C
78.2% -- 94.2% -- 387.4 #277
HRNet-W32-C
78.4% -- 94.2% -- 378.9 #254
HRNet-W40-C
78.9% -- 94.5% -- 355.8 #224
HRNet-W44-C
78.9% -- 94.4% -- 336.9 #227
HRNet-W48-C
79.3% -- 94.5% -- 315.7 #192
HRNet-W64-C
79.5% -- 94.7% -- 265.4 #182
Inception ResNet V2
80.5% 80.1%
95.3% 95.1% 287.1 #124
Inception V3
78.8% 78.8% 93.6% 94.4% 294.2 #233
Inception V4
80.2% -- 95.0% -- 288.0 #149
MixNet-L
79.0% 78.9% 94.0% 94.2% 349.3 #222
MixNet-M
77.0% 77.0% 93.4% 93.3% 357.4 #337
MixNet-S
75.6% 75.8% 92.8% 92.8% 360.3 #368
MixNet-XL
80.5% -- 94.9% -- 355.3 #153
MnasNet-A1
75.5% 75.2% 92.6% 92.5% 356.2 #373
MnasNet-B1
74.7% -- 92.1% -- 356.9 #401
MobileNet V3-Large 0.75
73.4% -- 91.4% -- 438.7 #430
MobileNet V3-Large 1.0
75.5% 75.2%
92.5% -- 430.1 #377
MobileNet V3-Large Minimal 1.0
72.2% -- 90.6% -- 410.5 #448
MobileNet V3-Small 0.75
65.7% -- 86.1% -- 434.5 #526
MobileNet V3-Small 1.0
67.9% -- 87.7% -- 440.4 #503
MobileNet V3-Small Minimal 1.0
62.9% -- 84.2% -- 414.4 #537
Modified Aligned Xception
79.6% 79.8% 94.7% 94.8% 260.0 #178
NASNet-A Large
82.6% -- 96.0% -- 116.7 #76
PNASNet-5
82.7% 82.9% 96.0% 96.2% 118.1 #71
RegNetX-12GF
79.6% -- 94.7% -- 339.5 #179
RegNetX-16GF
79.9% -- 94.8% -- 326.6 #164
RegNetX-1.6GF
76.9% -- 93.4% -- 367.3 #333
RegNetX-200MF
68.8% -- 88.5% -- 380.0 #495
RegNetX-32GF
80.2% -- 95.0% -- 174.5 #144
RegNetX-3.2GF
78.2% -- 94.1% -- 362.7 #269
RegNetX-400MF
72.4% -- 90.8% -- 371.2 #442
RegNetX-4.0GF
78.5% -- 94.2% -- 367.3 #255
RegNetX-600MF
73.9% -- 91.7% -- 372.5 #423
RegNetX-6.4GF
79.1% -- 94.5% -- 353.6 #208
RegNetX-800MF
75.0% -- 92.3% -- 377.7 #392
RegNetX-8.0GF
79.2% -- 94.6% -- 339.6 #195
RegNetY-12GF
80.4% -- 95.1% -- 344.3 #139
RegNetY-16GF
80.3% -- 95.0% -- 315.6 #151
RegNetY-1.6GF
77.9% -- 93.7% -- 372.9 #303
RegNetY-200MF
70.3% -- 89.5% -- 374.5 #473
RegNetY-32GF
80.8% -- 95.2% -- 183.5 #132
RegNetY-3.2GF
78.9% -- 94.4% -- 362.0 #231
RegNetY-400MF
74.0% -- 91.7% -- 380.4 #419
RegNetY-4.0GF
79.2% -- 94.7% -- 354.0 #200
RegNetY-600MF
75.3% -- 92.5% -- 372.7 #379
RegNetY-6.4GF
79.7% -- 94.8% -- 357.6 #173
RegNetY-800MF
76.3% -- 93.1% -- 375.0 #359
RegNetY-8.0GF
79.9% -- 94.8% -- 345.8 #163
Res2Net-50 14x8s
78.2% -- 93.8% -- 386.0 #292
Res2Net-50 26x4s
77.9% -- 93.9% -- 394.0 #291
Res2Net-50 26x6s
78.6% -- 94.1% -- 382.9 #249
Res2Net-50 26x8s
79.2% -- 94.4% -- 377.1 #231
Res2Net-50 48x2s
77.5% -- 93.5% -- 386.9 #311
Res2Net-DLA-60
78.5% 79.5% 94.2% -- 391.6 #257
Res2NeXt-101 26x4s
79.2% -- 94.4% -- 377.8 #210
Res2NeXt-50
78.2% -- 93.9% -- 377.0 #271
Res2NeXt-DLA-60
78.4% -- 94.1% -- 377.1 #259
ResNeSt-101
82.9% -- 96.3% -- 347.7 #72
ResNeSt-14
75.5% -- 92.5% -- 395.9 #375
ResNeSt-200
83.9% -- 96.9% -- 140.2 #51
ResNeSt-26
78.5% -- 94.3% -- 415.9 #238
ResNeSt-269
84.5% -- 97.0% -- 64.9 #41
ResNeSt-50
81.0% -- 95.4% -- 407.7 #118
ResNeSt-50 1s4x24d
81.0% -- 95.3% -- 349.3 #121
ResNeSt-50 4s2x40d
81.1% -- 95.6% -- 352.4 #111
ResNet-101
79.3% -- 94.5% -- 346.2 #194
ResNet-101-C
79.5% -- 94.6% -- 349.5 #180
ResNet-101-D
80.4% -- 95.0% -- 346.8 #145
ResNet-101-S
80.3% -- 95.2% -- 341.9 #145
ResNet-152
79.7% -- 94.7% -- 339.9 #175
ResNet-152-C
79.9% -- 94.8% -- 333.1 #161
ResNet-152-D
80.5% -- 95.2% -- 337.3 #134
ResNet-152-S
81.0% -- 95.4% -- 335.2 #114
ResNet-18
70.8% -- 89.1% -- 415.7 #468
ResNet-18
73.3% -- 91.4% -- 423.6 #431
ResNet-26
75.3% -- 92.6% -- 354.9 #374
ResNet-26-D
76.7% -- 93.2% -- 352.7 #352
ResNet-34
75.1% -- 92.3% -- 416.9 #389
ResNet-50
77.6% -- 93.7% -- 355.0 #309
ResNet-50
79.2% -- 94.8% -- 398.5 #199
ResNet-50
(288x288 Mean-Max Pooling)
82.0% -- 96.4% -- 343.8 #87
ResNet-50
(288x288 Mean-Max Pooling)
80.1% -- 95.2% -- 315.9 #130
ResNet-50-C
78.0% -- 94.0% -- 362.6 #278
ResNet-50-D
79.1% 77.2%
94.5% 93.5%
357.3 #204
ResNet-50-S
78.7% -- 94.2% -- 352.4 #242
ResNet-Blur-50
79.3% -- 94.6% -- 330.5 #187
ResNeXt-101 32x16d
84.2% -- 97.2% -- 150.9 #42
ResNeXt-101 32x16d
81.8% -- 96.9% -- 150.9 #90
ResNeXt-101 32x16d
(288x288 Mean-Max Pooling)
84.3% -- 97.3% -- 81.6 #40
ResNeXt-101 32x16d
(288x288 Mean-Max Pooling)
85.0% -- 97.6% -- 84.1 #27
ResNeXt-101 32x32d
85.1% 85.1% 97.4% 97.5% 60.5 #24
ResNeXt-101 32x32d
(288x288 Mean-Max Pooling)
85.9% -- 97.8% -- 37.6 #12
ResNeXt-101 32x48d
85.4% 85.4% 97.6% 97.6% 31.6 #15
ResNeXt-101 32x48d
(288x288 Mean-Max Pooling)
86.1% -- 97.9% -- 19.1 #9
ResNeXt-101 32x4d
80.3% -- 94.9% -- 342.0 #141
ResNeXt-101 32x4d
80.9% -- 96.8% -- 372.3 #56
ResNeXt-101 32x4d
(288x288 Mean-Max Pooling)
82.0% -- 96.3% -- 283.3 #84
ResNeXt-101 32x8d
84.3% -- 97.2% -- 286.4 #38
ResNeXt-101 32x8d
82.7% 82.2%
96.6% 96.4% 286.2 #63
ResNeXt-101 32x8d
(288x288 Mean-Max Pooling)
85.1% -- 97.6% -- 175.3 #23
ResNeXt-101 32x8d
(288x288 Mean-Max Pooling)
83.5% -- 97.1% -- 174.6 #40
ResNeXt-101 64x4d
80.6% -- 95.0% -- 277.5 #148
ResNeXt-50 32x4d
80.3% -- 95.4% -- 386.7 #142
ResNeXt-50 32x4d
79.8% -- 94.4% -- 352.2 #214
ResNeXt-50 32x4d
(288x288 Mean-Max Pooling)
81.3% -- 96.0% -- 341.8 #107
ResNeXt-50-D 32x4d
79.7% -- 94.9% -- 348.5 #160
SelecSLS-42_B
77.2% -- 93.4% -- 350.3 #337
SelecSLS-60
78.0% -- 93.8% -- 356.6 #294
SelecSLS-60_B
78.4% -- 94.2% -- 351.3 #261
SENet-154
81.3% 82.7% 95.5% 96.2% 215.6 #111
SENet-154
81.2% -- 95.4% -- 214.8 #109
SE-ResNet-101
78.4% -- 94.3% -- 391.5 #264
SE-ResNet-152
78.7% -- 94.4% -- 371.8 #245
SE-ResNet-18
71.8% -- 90.3% -- 354.7 #457
SE-ResNet-34
74.8% -- 92.1% -- 410.7 #400
SE-ResNet-50
77.6% -- 93.8% -- 399.1 #303
SE-ResNeXt-101 32x4d
80.9% -- 95.3% -- 335.2 #126
SE-ResNeXt-101 32x4d
80.2% -- 95.0% -- 375.5 #148
SE-ResNeXt-101 64x4d
80.9% -- 95.3% -- 266.0 #120
SE-ResNeXt-26 32x4d
77.1% -- 93.3% -- 353.9 #328
SE-ResNeXt-26-D 32x4d
77.6% -- 93.6% -- #311
SE-ResNeXt-26-T 32x4d
78.0% -- 93.7% -- #293
SE-ResNeXt-26-TN 32x4d
78.0% -- 93.7% -- 356.6 #298
SE-ResNeXt-50 32x4d
79.1% -- 94.4% -- 393.5 #212
SE-ResNeXt-50 32x4d
79.9% -- 94.8% -- 347.1 #167
Single-Path NAS
74.1% 75.0% 91.8% 92.2% 417.0 #411
SKNet-50
80.2% -- 94.6% -- 338.8 #185
SK-ResNet-18
73.0% -- 91.2% -- 346.4 #435
SK-ResNet-34
76.9% -- 93.3% -- 325.9 #342
VoVNet-39-V2
79.3% -- 94.7% -- 362.0 #177
Xception
79.0% 79.0% 94.4% 94.5% 295.6 #218
See Full Build Details +get badge code
[![SotaBench](https://img.shields.io/endpoint.svg?url=https://sotabench.com/api/v0/badge/gh/rwightman/pytorch-image-models)](https://sotabench.com/user/rwightman/repos/rwightman/pytorch-image-models)

How the Repository is Evaluated

The full sotabench.py file - source
import torch
from torchbench.image_classification import ImageNet
from timm import create_model
from timm.data import resolve_data_config, create_transform
from timm.models import TestTimePoolHead
import os

NUM_GPU = 1
BATCH_SIZE = 256 * NUM_GPU


def _entry(model_name, paper_model_name, paper_arxiv_id, batch_size=BATCH_SIZE,
           ttp=False, args=dict(), model_desc=None):
    return dict(
        model=model_name,
        model_description=model_desc,
        paper_model_name=paper_model_name,
        paper_arxiv_id=paper_arxiv_id,
        batch_size=batch_size,
        ttp=ttp,
        args=args)

# NOTE For any original PyTorch models, I'll remove from this list when you add to sotabench to
# avoid overlap and confusion. Please contact me.
model_list = [
    ## Weights ported by myself from other frameworks or trained myself in PyTorch
    _entry('adv_inception_v3', 'Adversarial Inception V3', '1611.01236',
           model_desc='Ported from official Tensorflow weights'),
    _entry('ens_adv_inception_resnet_v2', 'Ensemble Adversarial Inception V3', '1705.07204',
           model_desc='Ported from official Tensorflow weights'),
    _entry('dpn68', 'DPN-68 (224x224)', '1707.01629'),
    _entry('dpn68b', 'DPN-68b (224x224)', '1707.01629'),
    _entry('dpn92', 'DPN-92 (224x224)', '1707.01629'),
    _entry('dpn98', 'DPN-98 (224x224)', '1707.01629'),
    _entry('dpn107', 'DPN-107 (224x224)', '1707.01629'),
    _entry('dpn131', 'DPN-131 (224x224)', '1707.01629'),
    _entry('dpn68', 'DPN-68 (320x320, Mean-Max Pooling)', '1707.01629', ttp=True, args=dict(img_size=320)),
    _entry('dpn68b', 'DPN-68b (320x320, Mean-Max Pooling)', '1707.01629', ttp=True, args=dict(img_size=320)),
    _entry('dpn92', 'DPN-92 (320x320, Mean-Max Pooling)', '1707.01629',
           ttp=True, args=dict(img_size=320), batch_size=BATCH_SIZE//2),
    _entry('dpn98', 'DPN-98 (320x320, Mean-Max Pooling)', '1707.01629',
           ttp=True, args=dict(img_size=320), batch_size=BATCH_SIZE//2),
    _entry('dpn107', 'DPN-107 (320x320, Mean-Max Pooling)', '1707.01629',
           ttp=True, args=dict(img_size=320), batch_size=BATCH_SIZE//4),
    _entry('dpn131', 'DPN-131 (320x320, Mean-Max Pooling)', '1707.01629',
           ttp=True, args=dict(img_size=320), batch_size=BATCH_SIZE//4),
    _entry('efficientnet_b0', 'EfficientNet-B0', '1905.11946'),
    _entry('efficientnet_b1', 'EfficientNet-B1', '1905.11946'),
    _entry('efficientnet_b2', 'EfficientNet-B2', '1905.11946',
           model_desc='Trained from scratch in PyTorch w/ RandAugment'),
    _entry('efficientnet_b2a', 'EfficientNet-B2 (288x288, 1.0 crop)', '1905.11946',
           model_desc='Trained from scratch in PyTorch w/ RandAugment'),
    _entry('efficientnet_b3', 'EfficientNet-B3', '1905.11946',
           model_desc='Trained from scratch in PyTorch w/ RandAugment'),
    _entry('efficientnet_b3a', 'EfficientNet-B3 (320x320, 1.0 crop)', '1905.11946',
           model_desc='Trained from scratch in PyTorch w/ RandAugment'),
    _entry('efficientnet_es', 'EfficientNet-EdgeTPU-S', '1905.11946',
           model_desc='Trained from scratch in PyTorch w/ RandAugment'),

    _entry('gluon_inception_v3', 'Inception V3', '1512.00567', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet18_v1b', 'ResNet-18', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet34_v1b', 'ResNet-34', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet50_v1b', 'ResNet-50', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet50_v1c', 'ResNet-50-C', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet50_v1d', 'ResNet-50-D', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet50_v1s', 'ResNet-50-S', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet101_v1b', 'ResNet-101', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet101_v1c', 'ResNet-101-C', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet101_v1d', 'ResNet-101-D', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet101_v1s', 'ResNet-101-S', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet152_v1b', 'ResNet-152', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet152_v1c', 'ResNet-152-C', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet152_v1d', 'ResNet-152-D', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnet152_v1s', 'ResNet-152-S', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnext50_32x4d', 'ResNeXt-50 32x4d', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnext101_32x4d', 'ResNeXt-101 32x4d', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_resnext101_64x4d', 'ResNeXt-101 64x4d', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_senet154', 'SENet-154', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_seresnext50_32x4d', 'SE-ResNeXt-50 32x4d', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_seresnext101_32x4d', 'SE-ResNeXt-101 32x4d', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_seresnext101_64x4d', 'SE-ResNeXt-101 64x4d', '1812.01187', model_desc='Ported from GluonCV Model Zoo'),
    _entry('gluon_xception65', 'Modified Aligned Xception', '1802.02611', batch_size=BATCH_SIZE//2,
           model_desc='Ported from GluonCV Model Zoo'),

    _entry('mixnet_xl', 'MixNet-XL', '1907.09595', model_desc="My own scaling beyond paper's MixNet Large"),
    _entry('mixnet_l', 'MixNet-L', '1907.09595'),
    _entry('mixnet_m', 'MixNet-M', '1907.09595'),
    _entry('mixnet_s', 'MixNet-S', '1907.09595'),

    _entry('fbnetc_100', 'FBNet-C', '1812.03443',
           model_desc='Trained in PyTorch with RMSProp, exponential LR decay'),
    _entry('mnasnet_100', 'MnasNet-B1', '1807.11626'),
    _entry('semnasnet_100', 'MnasNet-A1', '1807.11626'),
    _entry('spnasnet_100', 'Single-Path NAS', '1904.02877',
           model_desc='Trained in PyTorch with SGD, cosine LR decay'),
    _entry('mobilenetv3_large_100', 'MobileNet V3-Large 1.0', '1905.02244',
           model_desc='Trained in PyTorch with RMSProp, exponential LR decay, and hyper-params matching '
                      'paper as closely as possible.'),

    _entry('resnet18', 'ResNet-18', '1812.01187'),
    _entry('resnet26', 'ResNet-26', '1812.01187', model_desc='Block cfg of ResNet-34 w/ Bottleneck'),
    _entry('resnet26d', 'ResNet-26-D', '1812.01187',
           model_desc='Block cfg of ResNet-34 w/ Bottleneck, deep stem, and avg-pool in downsample layers.'),
    _entry('resnet34', 'ResNet-34', '1812.01187'),
    _entry('resnet50', 'ResNet-50', '1812.01187', model_desc='Trained with AugMix + JSD loss'),
    _entry('resnet50', 'ResNet-50 (288x288 Mean-Max Pooling)', '1812.01187',
           ttp=True, args=dict(img_size=288),
           model_desc='Trained with AugMix + JSD loss'),
    _entry('resnext50_32x4d', 'ResNeXt-50 32x4d', '1812.01187'),
    _entry('resnext50d_32x4d', 'ResNeXt-50-D 32x4d', '1812.01187',
           model_desc="'D' variant (3x3 deep stem w/ avg-pool downscale). Trained with "
                      "SGD w/ cosine LR decay, random-erasing (gaussian per-pixel noise) and label-smoothing"),

    _entry('seresnet18', 'SE-ResNet-18', '1709.01507'),
    _entry('seresnet34', 'SE-ResNet-34', '1709.01507'),
    _entry('seresnext26_32x4d', 'SE-ResNeXt-26 32x4d', '1709.01507',
           model_desc='Block cfg of SE-ResNeXt-34 w/ Bottleneck'),
    _entry('seresnext26d_32x4d', 'SE-ResNeXt-26-D 32x4d', '1812.01187',
           model_desc='Block cfg of SE-ResNeXt-34 w/ Bottleneck, deep stem, and avg-pool in downsample layers.'),
    _entry('seresnext26t_32x4d', 'SE-ResNeXt-26-T 32x4d', '1812.01187',
           model_desc='Block cfg of SE-ResNeXt-34 w/ Bottleneck, deep tiered stem, and avg-pool in downsample layers.'),
    _entry('seresnext26tn_32x4d', 'SE-ResNeXt-26-TN 32x4d', '1812.01187',
           model_desc='Block cfg of SE-ResNeXt-34 w/ Bottleneck, deep tiered narrow stem, and avg-pool in downsample layers.'),

    _entry('skresnet18', 'SK-ResNet-18', '1903.06586'),
    _entry('skresnet34', 'SK-ResNet-34', '1903.06586'),
    _entry('skresnext50_32x4d', 'SKNet-50', '1903.06586'),

    _entry('ecaresnetlight', 'ECA-ResNet-Light', '1910.03151',
           model_desc='A tweaked ResNet50d with ECA attn.'),
    _entry('ecaresnet50d', 'ECA-ResNet-50d', '1910.03151',
           model_desc='A ResNet50d with ECA attn'),
    _entry('ecaresnet101d', 'ECA-ResNet-101d', '1910.03151',
           model_desc='A ResNet101d with ECA attn'),

    _entry('resnetblur50', 'ResNet-Blur-50', '1904.11486'),

    _entry('densenet121', 'DenseNet-121', '1608.06993'),
    _entry('densenetblur121d', 'DenseNet-Blur-121D', '1904.11486',
           model_desc='DenseNet with blur pooling and deep stem'),

    _entry('ese_vovnet39b', 'VoVNet-39-V2', '1911.06667'),

    _entry('tf_efficientnet_b0', 'EfficientNet-B0 (AutoAugment)', '1905.11946',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b1', 'EfficientNet-B1 (AutoAugment)', '1905.11946',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b2', 'EfficientNet-B2 (AutoAugment)', '1905.11946',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b3', 'EfficientNet-B3 (AutoAugment)', '1905.11946', batch_size=BATCH_SIZE//2,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b4', 'EfficientNet-B4 (AutoAugment)', '1905.11946', batch_size=BATCH_SIZE//2,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b5', 'EfficientNet-B5 (RandAugment)', '1905.11946', batch_size=BATCH_SIZE//4,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b6', 'EfficientNet-B6 (AutoAugment)', '1905.11946', batch_size=BATCH_SIZE//8,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b7', 'EfficientNet-B7 (RandAugment)', '1905.11946', batch_size=BATCH_SIZE//8,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b8', 'EfficientNet-B8 (RandAugment)', '1905.11946', batch_size=BATCH_SIZE // 8,
           model_desc='Ported from official Google AI Tensorflow weights'),

    _entry('tf_efficientnet_b0_ap', 'EfficientNet-B0 (AdvProp)', '1911.09665',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b1_ap', 'EfficientNet-B1 (AdvProp)', '1911.09665',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b2_ap', 'EfficientNet-B2 (AdvProp)', '1911.09665',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b3_ap', 'EfficientNet-B3 (AdvProp)', '1911.09665', batch_size=BATCH_SIZE // 2,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b4_ap', 'EfficientNet-B4 (AdvProp)', '1911.09665', batch_size=BATCH_SIZE // 2,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b5_ap', 'EfficientNet-B5 (AdvProp)', '1911.09665', batch_size=BATCH_SIZE // 4,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b6_ap', 'EfficientNet-B6 (AdvProp)', '1911.09665', batch_size=BATCH_SIZE // 8,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b7_ap', 'EfficientNet-B7 (AdvProp)', '1911.09665', batch_size=BATCH_SIZE // 8,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b8_ap', 'EfficientNet-B8 (AdvProp)', '1911.09665', batch_size=BATCH_SIZE // 8,
           model_desc='Ported from official Google AI Tensorflow weights'),

    _entry('tf_efficientnet_b0_ns', 'EfficientNet-B0 (NoisyStudent)', '1911.04252',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b1_ns', 'EfficientNet-B1 (NoisyStudent)', '1911.04252',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b2_ns', 'EfficientNet-B2 (NoisyStudent)', '1911.04252',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b3_ns', 'EfficientNet-B3 (NoisyStudent)', '1911.04252', batch_size=BATCH_SIZE // 2,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b4_ns', 'EfficientNet-B4 (NoisyStudent)', '1911.04252', batch_size=BATCH_SIZE // 2,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b5_ns', 'EfficientNet-B5 (NoisyStudent)', '1911.04252', batch_size=BATCH_SIZE // 4,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b6_ns', 'EfficientNet-B6 (NoisyStudent)', '1911.04252', batch_size=BATCH_SIZE // 8,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_b7_ns', 'EfficientNet-B7 (NoisyStudent)', '1911.04252', batch_size=BATCH_SIZE // 8,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_l2_ns_475', 'EfficientNet-L2 475 (NoisyStudent)', '1911.04252', batch_size=BATCH_SIZE // 16,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_l2_ns', 'EfficientNet-L2 (NoisyStudent)', '1911.04252', batch_size=BATCH_SIZE // 64,
           model_desc='Ported from official Google AI Tensorflow weights'),

    _entry('tf_efficientnet_cc_b0_4e', 'EfficientNet-CondConv-B0 4 experts', '1904.04971',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_cc_b0_8e', 'EfficientNet-CondConv-B0 8 experts', '1904.04971',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_cc_b1_8e', 'EfficientNet-CondConv-B1 8 experts', '1904.04971',
           model_desc='Ported from official Google AI Tensorflow weights'),

    _entry('tf_efficientnet_es', 'EfficientNet-EdgeTPU-S', '1905.11946',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_em', 'EfficientNet-EdgeTPU-M', '1905.11946',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_el', 'EfficientNet-EdgeTPU-L', '1905.11946', batch_size=BATCH_SIZE//2,
           model_desc='Ported from official Google AI Tensorflow weights'),

    _entry('tf_efficientnet_lite0', 'EfficientNet-Lite0', '1905.11946',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_lite1', 'EfficientNet-Lite1', '1905.11946',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_lite2', 'EfficientNet-Lite2', '1905.11946',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_lite3', 'EfficientNet-Lite3', '1905.11946', batch_size=BATCH_SIZE // 2,
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_efficientnet_lite4', 'EfficientNet-Lite4', '1905.11946', batch_size=BATCH_SIZE // 2,
           model_desc='Ported from official Google AI Tensorflow weights'),

    _entry('tf_inception_v3', 'Inception V3', '1512.00567', model_desc='Ported from official Tensorflow weights'),
    _entry('tf_mixnet_l', 'MixNet-L', '1907.09595', model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_mixnet_m', 'MixNet-M', '1907.09595', model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_mixnet_s', 'MixNet-S', '1907.09595', model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_mobilenetv3_large_100', 'MobileNet V3-Large 1.0', '1905.02244',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_mobilenetv3_large_075', 'MobileNet V3-Large 0.75', '1905.02244',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_mobilenetv3_large_minimal_100', 'MobileNet V3-Large Minimal 1.0', '1905.02244',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_mobilenetv3_small_100', 'MobileNet V3-Small 1.0', '1905.02244',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_mobilenetv3_small_075', 'MobileNet V3-Small 0.75', '1905.02244',
           model_desc='Ported from official Google AI Tensorflow weights'),
    _entry('tf_mobilenetv3_small_minimal_100', 'MobileNet V3-Small Minimal 1.0', '1905.02244',
           model_desc='Ported from official Google AI Tensorflow weights'),

    ## Cadene ported weights (to remove if Cadene adds sotabench)
    _entry('inception_resnet_v2', 'Inception ResNet V2', '1602.07261'),
    _entry('inception_v4', 'Inception V4', '1602.07261'),
    _entry('nasnetalarge', 'NASNet-A Large', '1707.07012', batch_size=BATCH_SIZE // 4),
    _entry('pnasnet5large', 'PNASNet-5', '1712.00559', batch_size=BATCH_SIZE // 4),
    _entry('seresnet50', 'SE-ResNet-50', '1709.01507'),
    _entry('seresnet101', 'SE-ResNet-101', '1709.01507'),
    _entry('seresnet152', 'SE-ResNet-152', '1709.01507'),
    _entry('seresnext50_32x4d', 'SE-ResNeXt-50 32x4d', '1709.01507'),
    _entry('seresnext101_32x4d', 'SE-ResNeXt-101 32x4d', '1709.01507'),
    _entry('senet154', 'SENet-154', '1709.01507'),
    _entry('xception', 'Xception', '1610.02357',  batch_size=BATCH_SIZE//2),

    ## Torchvision weights
    # _entry('densenet121'),
    # _entry('densenet161'),
    # _entry('densenet169'),
    # _entry('densenet201'),
    # _entry('inception_v3', paper_model_name='Inception V3', ),
    # _entry('tv_resnet34', , ),
    # _entry('tv_resnet50', , ),
    # _entry('resnet101', , ),
    # _entry('resnet152', , ),
    # _entry('tv_resnext50_32x4d', , ),
    # _entry('resnext101_32x8d', ),
    # _entry('wide_resnet50_2' , ),
    # _entry('wide_resnet101_2', , ),

    ## Facebook WSL weights
    _entry('ig_resnext101_32x8d', 'ResNeXt-101 32x8d', '1805.00932',
           model_desc='Weakly-Supervised pre-training on 1B Instagram hashtag dataset by Facebook Research'),
    _entry('ig_resnext101_32x16d', 'ResNeXt-101 32x16d', '1805.00932',
           model_desc='Weakly-Supervised pre-training on 1B Instagram hashtag dataset by Facebook Research'),
    _entry('ig_resnext101_32x32d', 'ResNeXt-101 32x32d', '1805.00932', batch_size=BATCH_SIZE // 2,
           model_desc='Weakly-Supervised pre-training on 1B Instagram hashtag dataset by Facebook Research'),
    _entry('ig_resnext101_32x48d', 'ResNeXt-101 32x48d', '1805.00932', batch_size=BATCH_SIZE // 4,
           model_desc='Weakly-Supervised pre-training on 1B Instagram hashtag dataset by Facebook Research'),

    _entry('ig_resnext101_32x8d', 'ResNeXt-101 32x8d (288x288 Mean-Max Pooling)', '1805.00932',
           ttp=True, args=dict(img_size=288),
           model_desc='Weakly-Supervised pre-training on 1B Instagram hashtag dataset by Facebook Research'),
    _entry('ig_resnext101_32x16d', 'ResNeXt-101 32x16d (288x288 Mean-Max Pooling)', '1805.00932',
           ttp=True, args=dict(img_size=288), batch_size=BATCH_SIZE // 2,
           model_desc='Weakly-Supervised pre-training on 1B Instagram hashtag dataset by Facebook Research'),
    _entry('ig_resnext101_32x32d', 'ResNeXt-101 32x32d (288x288 Mean-Max Pooling)', '1805.00932',
           ttp=True, args=dict(img_size=288), batch_size=BATCH_SIZE // 4,
           model_desc='Weakly-Supervised pre-training on 1B Instagram hashtag dataset by Facebook Research'),
    _entry('ig_resnext101_32x48d', 'ResNeXt-101 32x48d (288x288 Mean-Max Pooling)', '1805.00932',
           ttp=True, args=dict(img_size=288), batch_size=BATCH_SIZE // 8,
           model_desc='Weakly-Supervised pre-training on 1B Instagram hashtag dataset by Facebook Research'),

    ## Facebook SSL weights
    _entry('ssl_resnet18', 'ResNet-18', '1905.00546',
           model_desc='Semi-Supervised pre-training on YFCC100M dataset by Facebook Research'),
    _entry('ssl_resnet50', 'ResNet-50', '1905.00546',
           model_desc='Semi-Supervised pre-training on YFCC100M dataset by Facebook Research'),
    _entry('ssl_resnext50_32x4d', 'ResNeXt-50 32x4d', '1905.00546',
           model_desc='Semi-Supervised pre-training on YFCC100M dataset by Facebook Research'),
    _entry('ssl_resnext101_32x4d', 'ResNeXt-101 32x4d', '1905.00546',
           model_desc='Semi-Supervised pre-training on YFCC100M dataset by Facebook Research'),
    _entry('ssl_resnext101_32x8d', 'ResNeXt-101 32x8d', '1905.00546',
           model_desc='Semi-Supervised pre-training on YFCC100M dataset by Facebook Research'),
    _entry('ssl_resnext101_32x16d', 'ResNeXt-101 32x16d', '1905.00546',
           model_desc='Semi-Supervised pre-training on YFCC100M dataset by Facebook Research'),

    _entry('ssl_resnet50', 'ResNet-50 (288x288 Mean-Max Pooling)', '1905.00546',
           ttp=True, args=dict(img_size=288),
           model_desc='Semi-Supervised pre-training on YFCC100M dataset by Facebook Research'),
    _entry('ssl_resnext50_32x4d', 'ResNeXt-50 32x4d (288x288 Mean-Max Pooling)', '1905.00546',
           ttp=True, args=dict(img_size=288),
           model_desc='Semi-Supervised pre-training on YFCC100M dataset by Facebook Research'),
    _entry('ssl_resnext101_32x4d', 'ResNeXt-101 32x4d (288x288 Mean-Max Pooling)', '1905.00546',
           ttp=True, args=dict(img_size=288),
           model_desc='Semi-Supervised pre-training on YFCC100M dataset by Facebook Research'),
    _entry('ssl_resnext101_32x8d', 'ResNeXt-101 32x8d (288x288 Mean-Max Pooling)', '1905.00546',
           ttp=True, args=dict(img_size=288),
           model_desc='Semi-Supervised pre-training on YFCC100M dataset by Facebook Research'),
    _entry('ssl_resnext101_32x16d', 'ResNeXt-101 32x16d (288x288 Mean-Max Pooling)', '1905.00546',
           ttp=True, args=dict(img_size=288), batch_size=BATCH_SIZE // 2,
           model_desc='Semi-Supervised pre-training on YFCC100M dataset by Facebook Research'),

    ## Facebook SWSL weights
    _entry('swsl_resnet18', 'ResNet-18', '1905.00546',
           model_desc='Semi-Weakly-Supervised pre-training on 1 billion unlabelled dataset by Facebook Research'),
    _entry('swsl_resnet50', 'ResNet-50', '1905.00546',
           model_desc='Semi-Weakly-Supervised pre-training on 1 billion unlabelled dataset by Facebook Research'),
    _entry('swsl_resnext50_32x4d', 'ResNeXt-50 32x4d', '1905.00546',
           model_desc='Semi-Weakly-Supervised pre-training on 1 billion unlabelled dataset by Facebook Research'),
    _entry('swsl_resnext101_32x4d', 'ResNeXt-101 32x4d', '1905.00546',
           model_desc='Semi-Weakly-Supervised pre-training on 1 billion unlabelled dataset by Facebook Research'),
    _entry('swsl_resnext101_32x8d', 'ResNeXt-101 32x8d', '1905.00546',
           model_desc='Semi-Weakly-Supervised pre-training on 1 billion unlabelled dataset by Facebook Research'),
    _entry('swsl_resnext101_32x16d', 'ResNeXt-101 32x16d', '1905.00546',
           model_desc='Semi-Weakly-Supervised pre-training on 1 billion unlabelled dataset by Facebook Research'),

    _entry('swsl_resnet50', 'ResNet-50 (288x288 Mean-Max Pooling)', '1905.00546',
           ttp=True, args=dict(img_size=288),
           model_desc='Semi-Weakly-Supervised pre-training on 1 billion unlabelled dataset by Facebook Research'),
    _entry('swsl_resnext50_32x4d', 'ResNeXt-50 32x4d (288x288 Mean-Max Pooling)', '1905.00546',
           ttp=True, args=dict(img_size=288),
           model_desc='Semi-Weakly-Supervised pre-training on 1 billion unlabelled dataset by Facebook Research'),
    _entry('swsl_resnext101_32x4d', 'ResNeXt-101 32x4d (288x288 Mean-Max Pooling)', '1905.00546',
           ttp=True, args=dict(img_size=288),
           model_desc='Semi-Weakly-Supervised pre-training on 1 billion unlabelled dataset by Facebook Research'),
    _entry('swsl_resnext101_32x8d', 'ResNeXt-101 32x8d (288x288 Mean-Max Pooling)', '1905.00546',
           ttp=True, args=dict(img_size=288),
           model_desc='Semi-Weakly-Supervised pre-training on 1 billion unlabelled dataset by Facebook Research'),
    _entry('swsl_resnext101_32x16d', 'ResNeXt-101 32x16d (288x288 Mean-Max Pooling)', '1905.00546',
           ttp=True, args=dict(img_size=288), batch_size=BATCH_SIZE // 2,
           model_desc='Semi-Weakly-Supervised pre-training on 1 billion unlabelled dataset by Facebook Research'),

    ## DLA official impl weights (to remove if sotabench added to source)
    _entry('dla34', 'DLA-34', '1707.06484'),
    _entry('dla46_c', 'DLA-46-C', '1707.06484'),
    _entry('dla46x_c', 'DLA-X-46-C', '1707.06484'),
    _entry('dla60x_c', 'DLA-X-60-C', '1707.06484'),
    _entry('dla60', 'DLA-60', '1707.06484'),
    _entry('dla60x', 'DLA-X-60', '1707.06484'),
    _entry('dla102', 'DLA-102', '1707.06484'),
    _entry('dla102x', 'DLA-X-102', '1707.06484'),
    _entry('dla102x2', 'DLA-X-102 64', '1707.06484'),
    _entry('dla169', 'DLA-169', '1707.06484'),

    ## Res2Net official impl weights (to remove if sotabench added to source)
    _entry('res2net50_26w_4s', 'Res2Net-50 26x4s', '1904.01169'),
    _entry('res2net50_14w_8s', 'Res2Net-50 14x8s', '1904.01169'),
    _entry('res2net50_26w_6s', 'Res2Net-50 26x6s', '1904.01169'),
    _entry('res2net50_26w_8s', 'Res2Net-50 26x8s', '1904.01169'),
    _entry('res2net50_48w_2s', 'Res2Net-50 48x2s', '1904.01169'),
    _entry('res2net101_26w_4s', 'Res2NeXt-101 26x4s', '1904.01169'),
    _entry('res2next50', 'Res2NeXt-50', '1904.01169'),
    _entry('dla60_res2net', 'Res2Net-DLA-60', '1904.01169'),
    _entry('dla60_res2next', 'Res2NeXt-DLA-60', '1904.01169'),

    ## HRNet official impl weights
    _entry('hrnet_w18_small', 'HRNet-W18-C-Small-V1', '1908.07919'),
    _entry('hrnet_w18_small_v2', 'HRNet-W18-C-Small-V2', '1908.07919'),
    _entry('hrnet_w18', 'HRNet-W18-C', '1908.07919'),
    _entry('hrnet_w30', 'HRNet-W30-C', '1908.07919'),
    _entry('hrnet_w32', 'HRNet-W32-C', '1908.07919'),
    _entry('hrnet_w40', 'HRNet-W40-C', '1908.07919'),
    _entry('hrnet_w44', 'HRNet-W44-C', '1908.07919'),
    _entry('hrnet_w48', 'HRNet-W48-C', '1908.07919'),
    _entry('hrnet_w64', 'HRNet-W64-C', '1908.07919'),


    ## SelecSLS official impl weights
    _entry('selecsls42b', 'SelecSLS-42_B', '1907.00837',
           model_desc='Originally from https://github.com/mehtadushy/SelecSLS-Pytorch'),
    _entry('selecsls60', 'SelecSLS-60', '1907.00837',
           model_desc='Originally from https://github.com/mehtadushy/SelecSLS-Pytorch'),
    _entry('selecsls60b', 'SelecSLS-60_B', '1907.00837',
           model_desc='Originally from https://github.com/mehtadushy/SelecSLS-Pytorch'),

    ## ResNeSt official impl weights
    _entry('resnest14d', 'ResNeSt-14', '2004.08955',
           model_desc='Originally from GluonCV'),
    _entry('resnest26d', 'ResNeSt-26', '2004.08955',
           model_desc='Originally from GluonCV'),
    _entry('resnest50d', 'ResNeSt-50', '2004.08955',
           model_desc='Originally from https://github.com/zhanghang1989/ResNeSt'),
    _entry('resnest101e', 'ResNeSt-101', '2004.08955',
           model_desc='Originally from https://github.com/zhanghang1989/ResNeSt'),
    _entry('resnest200e', 'ResNeSt-200', '2004.08955',
           model_desc='Originally from https://github.com/zhanghang1989/ResNeSt'),
    _entry('resnest269e', 'ResNeSt-269', '2004.08955', batch_size=BATCH_SIZE // 2,
           model_desc='Originally from https://github.com/zhanghang1989/ResNeSt'),
    _entry('resnest50d_4s2x40d', 'ResNeSt-50 4s2x40d', '2004.08955',
           model_desc='Originally from https://github.com/zhanghang1989/ResNeSt'),
    _entry('resnest50d_1s4x24d', 'ResNeSt-50 1s4x24d', '2004.08955',
           model_desc='Originally from https://github.com/zhanghang1989/ResNeSt'),

    ## RegNet official impl weighs
    _entry('regnetx_002', 'RegNetX-200MF', '2003.13678'),
    _entry('regnetx_004', 'RegNetX-400MF', '2003.13678'),
    _entry('regnetx_006', 'RegNetX-600MF', '2003.13678'),
    _entry('regnetx_008', 'RegNetX-800MF', '2003.13678'),
    _entry('regnetx_016', 'RegNetX-1.6GF', '2003.13678'),
    _entry('regnetx_032', 'RegNetX-3.2GF', '2003.13678'),
    _entry('regnetx_040', 'RegNetX-4.0GF', '2003.13678'),
    _entry('regnetx_064', 'RegNetX-6.4GF', '2003.13678'),
    _entry('regnetx_080', 'RegNetX-8.0GF', '2003.13678'),
    _entry('regnetx_120', 'RegNetX-12GF', '2003.13678'),
    _entry('regnetx_160', 'RegNetX-16GF', '2003.13678'),
    _entry('regnetx_320', 'RegNetX-32GF', '2003.13678', batch_size=BATCH_SIZE // 2),

    _entry('regnety_002', 'RegNetY-200MF', '2003.13678'),
    _entry('regnety_004', 'RegNetY-400MF', '2003.13678'),
    _entry('regnety_006', 'RegNetY-600MF', '2003.13678'),
    _entry('regnety_008', 'RegNetY-800MF', '2003.13678'),
    _entry('regnety_016', 'RegNetY-1.6GF', '2003.13678'),
    _entry('regnety_032', 'RegNetY-3.2GF', '2003.13678'),
    _entry('regnety_040', 'RegNetY-4.0GF', '2003.13678'),
    _entry('regnety_064', 'RegNetY-6.4GF', '2003.13678'),
    _entry('regnety_080', 'RegNetY-8.0GF', '2003.13678'),
    _entry('regnety_120', 'RegNetY-12GF', '2003.13678'),
    _entry('regnety_160', 'RegNetY-16GF', '2003.13678'),
    _entry('regnety_320', 'RegNetY-32GF', '2003.13678', batch_size=BATCH_SIZE // 2),

]

for m in model_list:
    model_name = m['model']
    # create model from name
    model = create_model(model_name, pretrained=True)
    param_count = sum([m.numel() for m in model.parameters()])
    print('Model %s, %s created. Param count: %d' % (model_name, m['paper_model_name'], param_count))

    # get appropriate transform for model's default pretrained config
    data_config = resolve_data_config(m['args'], model=model, verbose=True)
    if m['ttp']:
        model = TestTimePoolHead(model, model.default_cfg['pool_size'])
        data_config['crop_pct'] = 1.0
    input_transform = create_transform(**data_config)

    # Run the benchmark
    ImageNet.benchmark(
        model=model,
        model_description=m.get('model_description', None),
        paper_model_name=m['paper_model_name'],
        paper_arxiv_id=m['paper_arxiv_id'],
        input_transform=input_transform,
        batch_size=m['batch_size'],
        num_gpu=NUM_GPU,
        data_root=os.environ.get('IMAGENET_DIR', './imagenet')
    )

    torch.cuda.empty_cache()


STATUS
BUILD
COMMIT MESSAGE
RUN TIME
Add `adamp` and 'sgdp' optimizers. Update requirements.txt Upd…
rwightman   e93e571  ·  Jul 25 2020
0h:11m:34s
Add autosquash workflow
rwightman   0915bed  ·  Jul 25 2020
0h:12m:37s
Merge branch 'michalwols-docs'
rwightman   17f4dd2  (+2 commits )  ·  Jul 10 2020
0h:14m:10s
Update setup.py Exclude results from possible packaging as it h…
rwightman   31cf125  ·  Jul 09 2020
0h:13m:45s
Merge pull request #183 from KushajveerSingh/results_diff Add r…
rwightman   0d5550c  (+3 commits )  ·  Jul 09 2020
0h:14m:36s
Fix #173, lr cycle default 0 vs 1. Fix #177, mirror resnest weig…
rwightman   d72ac0d  ·  Jun 29 2020
0h:49m:04s
Remove tests from distrib
rwightman   24e7535  ·  Jun 16 2020
0h:52m:04s
Add ESE-VoVNet-19-DW weights
rwightman   328339a  ·  Jun 15 2020
0h:48m:00s
Fix default interpolation/crop of largest 2 ResNeSt models
rwightman   2d83752  ·  Jun 12 2020
1h:03m:43s
Update README with model results and attribution. Make scheduler…
rwightman   f225ae8  ·  Jun 12 2020
1h:20m:54s
Merge pull request #155 from rwightman/densenet_update_and_more …
rwightman   d1b5ddd  (+22 commits )  ·  Jun 11 2020
2h:00m:55s
Update requirements so PyTorch 1.4 is min, add separate sotabenc…
rwightman   5966654  ·  May 24 2020
0h:40m:56s
Update sotabench.py
rwightman   d79ac48  ·  May 22 2020
0h:12m:19s
Merge pull request #154 from rwightman/tests_bugfixes Add backw…
rwightman   e881383  (+5 commits )  ·  May 21 2020
0h:11m:09s
Merge pull request #150 from rwightman/regnet Add RegNet models…
rwightman 50658b9    ea2e59c  ·  May 18 2020
0h:11m:47s
Merge pull request #148 from rwightman/drop_block_improve Impro…
rwightman 1904ed8    dab9935  ·  May 13 2020
0h:11m:37s
Merge pull request #146 from rwightman/inceptionv3_fix Remove a…
rwightman 17270c6    63addb7  ·  May 12 2020
0h:12m:50s
Merge pull request #145 from rwightman/resnest ResNeSt
rwightman   c4ca016  (+3 commits )  ·  May 12 2020
0h:14m:19s
Refactor test indent
rwightman 9cc289f    5bd1ad1  ·  May 12 2020
0h:16m:12s
Update test_inference.py Not six min
rwightman   e545bb9  ·  May 07 2020
0h:09m:36s
Update test_inference.py Make the timeout 5-min for now, see if…
rwightman   305a2db  ·  May 07 2020
0h:18m:33s
Merge pull request #143 from michalwols/master Setup Github Act…
rwightman   14e01b8  (+4 commits )  ·  May 07 2020
0h:19m:11s
Merge pull request #141 from Animatory/fix_HRNet Fixed HRNet mo…
rwightman f0eb021    6cc11a8  ·  May 05 2020
0h:15m:23s
Merge pull request #140 from yoniaflalo/PR_MultiEpochsDataLoader…
rwightman a7f570c    3b72ebf  ·  May 05 2020
0h:17m:04s
0h:14m:54s
Fix #139. Broken SKResNets after BlurPool addition, as a plus, S…
rwightman   8d8677e  ·  May 04 2020
0h:16m:12s
Update README.md
rwightman   353a79a  ·  May 03 2020
0h:13m:54s
Bump version for Pypi release
rwightman   c9b6f41  ·  May 03 2020
0h:16m:30s
Add EfficientNet pruned models to results files
rwightman   31f4c12  ·  May 03 2020
0h:16m:33s
Fix pruned txt files not being installed during pip install
rwightman   375f3e5  ·  May 03 2020
0h:17m:53s
Merge pull request #136 from yoniaflalo/adding_effnet_pruned ad…
rwightman 9c15d57    8ec554b  ·  May 03 2020
0h:12m:04s
Update README.md
rwightman   a4d20a1  ·  May 01 2020
0h:09m:39s
Fix model create fn not passing num_classes through. Fix #135
rwightman   ea30070  ·  May 01 2020
0h:16m:44s
Update results with new models
rwightman   779cb0f  ·  May 01 2020
0h:16m:45s
Merge branch 'master' of github.com:rwightman/pytorch-models
rwightman   2c438c4  (+9 commits )  ·  May 01 2020
0h:15m:56s
Merge pull request #125 from Separius/patch-1 fix typo in eca
rwightman 20290b5    a5220ad  ·  May 01 2020
0h:19m:08s
Merge branch 'yoniaflalo-adding_ECA_resnet'
rwightman   7a9942a  (+3 commits )  ·  May 01 2020
0h:17m:11s
Bump version for pypi release. Fix #130
rwightman   1d4ac1b  (+6 commits )  ·  Apr 27 2020
0h:15m:31s
Merge pull request #122 from mrT23/master TResNet models
rwightman   ebf82b8  (+3 commits )  ·  Apr 12 2020
0h:49m:15s
Merge pull request #123 from aclex/mobilenetv3_fix_feature_extra…
rwightman bdb165a    e15f979  ·  Apr 12 2020
0h:51m:55s
Remove poorly named metrics from torch imagenet example origins.…
rwightman   13cf688  ·  Apr 10 2020
0h:52m:25s
Bump version for pypi
rwightman   56608c9  (+2 commits )  ·  Apr 09 2020
0h:50m:09s
Merge pull request #117 from VRandme/typo_eca minor PR to fix t…
rwightman 06a50a9    e01ccb8  ·  Apr 07 2020
0h:50m:19s
Merge pull request #115 from rwightman/mobilenetv2-experiment M…
rwightman   c99a5ab  (+3 commits )  ·  Apr 05 2020
0h:50m:19s
Add better resnext50_32x4d weights trained by andravin
rwightman   5a16c53  ·  Mar 18 2020
0h:52m:01s
Merge pull request #105 from rwightman/efficientnet-lite Effici…
rwightman   71b5cd6  (+3 commits )  ·  Mar 18 2020
1h:04m:08s
Merge pull request #99 from andravin/save-last Modified save_ch…
rwightman d92cc4d    7deacf5  ·  Mar 15 2020
0h:47m:33s
Merge pull request #94 from rwightman/lr_noise Learning rate no…
rwightman   56e2ac3  (+4 commits )  ·  Feb 29 2020
0h:52m:50s
Annotate types on drop fns to avoid torchscript error
rwightman   c60069c  ·  Feb 27 2020
0h:52m:10s
version bump for PyPi update
rwightman   cc5a11a  ·  Feb 22 2020
0h:48m:15s
Forgot to add skresnet34 to sotabench
rwightman   d77f45a  ·  Feb 18 2020
0h:48m:25s
Simpler approach to loading entrypoints in hubconf works properly
rwightman   6620770  ·  Feb 18 2020
unknown
Merge pull request #88 from rwightman/attention A lot of attent…
rwightman   e0685dd  (+42 commits )  ·  Feb 18 2020
3h:03m:51s
Add map_location='cpu' to ModelEma resume, should improve #72
rwightman   f098fda  ·  Feb 12 2020
0h:47m:23s
Add L2-475 PyTorch preprocessing result, update sotabench for ne…
rwightman   b949699  ·  Feb 12 2020
4h:27m:05s
Add ported EfficientNet-L2, B0-B7 NoisyStudent weights from TF T…
rwightman   ba15ca4  ·  Feb 12 2020
0h:47m:19s
Remove unused default_init for EfficientNets, experimenting with…
rwightman cade829    d0eb59e  ·  Feb 09 2020
0h:43m:16s
Update README.md
rwightman   5eb0e36  ·  Feb 06 2020
0h:48m:20s
Add PyTorch trained EfficientNet-ES weights from Andrew Lavin
rwightman   5c4991a  ·  Feb 06 2020
0h:44m:26s
Indentation mistake. Fixes #81
rwightman   d66819d  ·  Feb 04 2020
0h:48m:40s
Merge pull request #83 from andravin/validation-batch-size-multi…
rwightman b72013d    65cda1c  ·  Feb 04 2020
0h:47m:33s
Bump version for PyPi update, fix few out of date README items/m…
rwightman   4808b3c  ·  Feb 03 2020
0h:46m:53s
Update README.md
rwightman   5c85389  ·  Feb 02 2020
0h:42m:13s
Update README.md
rwightman   820b73d  ·  Feb 02 2020
unknown
Update README.md Fix relative paths (I think)
rwightman   82c0a2f  ·  Feb 02 2020
unknown
Add results/README.md
rwightman   1ffd2d0  ·  Feb 02 2020
0h:48m:03s
Update sotabench with tf_efficientnet_b8 model
rwightman   fd98fb3  (+4 commits )  ·  Feb 01 2020
1h:12m:15s
Add warning about using sync-bn with zero initialized BN layers.…
rwightman   5b7cc16  ·  Jan 31 2020
0h:41m:02s
Update README.md Typo
rwightman   b18c199  ·  Jan 31 2020
unknown
New ResNet50 JSD + RandAugment weights
rwightman   12dbc74  ·  Jan 31 2020
0h:51m:06s
0h:55m:04s
Update README.md Fix typo
rwightman   e39aae5  ·  Jan 13 2020
0h:05m:21s
Update README.md
rwightman   7a17ee9  (+3 commits )  ·  Jan 12 2020
0h:09m:57s
Update README.md
rwightman   2a88412  ·  Jan 11 2020
0h:05m:23s
Merge pull request #74 from rwightman/augmix-jsd AugMix, JSD lo…
rwightman   d9a6a9d  (+7 commits )  ·  Jan 11 2020
0h:08m:27s
Add tiered narrow ResNet (tn) and weights for seresnext26tn_32x4d
rwightman   a28117e  ·  Jan 11 2020
0h:09m:09s
Update README.md
rwightman   cfa951b  ·  Jan 07 2020
0h:09m:41s
Update README.md
rwightman   7622015  ·  Jan 04 2020
0h:09m:11s
Add updated RandAugment trained EfficientNet-B0 trained weights …
rwightman   ec0dd40  ·  Jan 03 2020
0h:10m:16s
Plural for averaging script.
rwightman   8662454  ·  Jan 03 2020
0h:10m:16s
Add checkpoint averaging script. Add headers, shebangs, exec per…
rwightman 4666cc9    40fea63  ·  Jan 03 2020
0h:11m:02s
ResNet / Res2Net additions: * ResNet torchscript compat * output…
rwightman   53001dd  (+2 commits )  ·  Jan 01 2020
0h:46m:03s
Update README.md Update readme with SelecSLS details.
rwightman   a7fe891  ·  Dec 30 2019
0h:46m:43s
Merge pull request #66 from rwightman/selecsls_updates SelecSLS…
rwightman   e728d70  (+4 commits )  ·  Dec 30 2019
0h:51m:07s
Merge pull request #65 from mehtadushy/selecsls Incorporate Sel…
rwightman   fb3a0f4  (+3 commits )  ·  Dec 30 2019
0h:42m:33s
Update README.md
rwightman   0554b79  ·  Dec 29 2019
0h:40m:38s
Update sotabench.py
rwightman   a4497af  ·  Dec 28 2019
0h:44m:51s
Update README with B3 training details
rwightman   53f578e  (+3 commits )  ·  Dec 28 2019
0h:56m:32s
Add ResNet deep tiered stem and model weights for seresnext26t_3…
rwightman   1f4498f  ·  Dec 28 2019
0h:45m:56s
Add update RandAugment MixNet-XL weights
rwightman   73b7845  ·  Dec 24 2019
0h:49m:50s
Merge pull request #62 from rwightman/reduce-bn Distribute Batc…
rwightman   ff8688c  (+4 commits )  ·  Dec 19 2019
0h:46m:20s
Update README.md Update latest training hparam/command line wit…
rwightman   5d7af97  ·  Dec 05 2019
0h:33m:56s
Update README.md
rwightman   3129bdb  ·  Dec 04 2019
0h:40m:02s
New PyTorch trained EfficientNet-B2 weights with my RandAugment …
rwightman   ff421e5  ·  Dec 04 2019
0h:39m:42s
Update results-all.csv with latest models/weights
rwightman   00b9340  ·  Nov 29 2019
0h:34m:00s
Update README.md with latest changes
rwightman   5259dbc  ·  Nov 29 2019
0h:35m:42s
Finish with HRNet, weights and models updated. Improve consisten…
rwightman   3bef524  ·  Nov 29 2019
0h:55m:33s
Merge pull request #53 from rwightman/condconvs_and_features Ma…
rwightman   3ceeedc  (+12 commits )  ·  Nov 28 2019
2h:01m:52s
Merge pull request #52 from rwightman/randaugment RandAugment a…
rwightman   db04677  (+5 commits )  ·  Nov 22 2019
0h:31m:20s
Fix non-prefetch variant of Mixup. Fixes #50
rwightman   4748c6d  ·  Nov 02 2019
0h:32m:07s
Add TF RandAug weights for B5/B7 EfficientNet models.
rwightman   0d58c50  ·  Oct 30 2019
0h:57m:20s
Better differentiate sotabench WSL, SSL, and SWSL models via mod…
rwightman   62105ed  ·  Oct 20 2019
0h:31m:09s
Map pretrained checkpoint to cpu to avoid issue with some pretra…
rwightman   c099374  ·  Oct 19 2019
0h:29m:28s
Add Facebook Research Semi-Supervised and Semi-Weakly Supervised…
rwightman a9eb484    b93fcf0  ·  Oct 19 2019
7h:41m:03s
6h:58m:34s
Add support for loading args from yaml file (and saving them wit…
rwightman   187ecba  ·  Sep 09 2019
0h:27m:02s
Fix Mobilenet V3 model name for sotabench. Minor res2net cleanup.
rwightman   d3ba34e  ·  Sep 05 2019
2h:26m:03s
Silly typos.
rwightman   b5a8bb5  ·  Sep 04 2019
3h:23m:37s
sotabench fail
rwightman   7dc5d7a  ·  Sep 04 2019
1h:11m:43s
Merge pull request #35 from rwightman/res2net_dla Add Res2net a…
rwightman   96364fc  (+2 commits )  ·  Sep 04 2019
0h:58m:37s
1h:36m:10s
0h:58m:34s