Leaderboard Explore Compare Profile Community Blog FAQ Competition 2024 Report Issue

Sample stimuli

How to use

from brainscore_vision import load_benchmark
benchmark = load_benchmark("ObjectNet-top1")
score = benchmark(my_model)

Benchmark API

Code examples

Model scores

Score Legend Min Alignment Max Alignment

Rank	Model	Score
1	cvt_cvt-w24-384-in22k_finetuned-in1k_4	.513
2	resnext101_32x48d_wsl	.499
3	fixres_resnext101_32x48d_wsl	.491
4	resnext101_32x8d_wsl	.443
5	convnext_tiny_sup	.365
6	focalnet_tiny_lrf_in1k	.346
7	nasnet_large-tf	.324
8	pnasnet_large-tf	.316
9	antialias-resnet152	.309
10	effnetb1_272x240	.305
11	inception_resnet_v2	.304
12	AdvProp_efficientnet-b7	.297
13	AdvProp_efficientnet-b6	.294
14	CLIP-RN50	.293
15	inception_v4-tf	.291
16	efficientnet-b7	.285
17	cv_18_dagger_408_pretrained	.277
18	antialiased-r50	.272
19	efficientnet_b0	.270
20	ViT_L_32_imagenet1k	.266
21	AdvProp_efficientnet-b2	.265
22	AT_efficientnet-b2	.265
23	AdvProp_efficientnet-b4	.265
24	AT_efficientnet-b4	.265
25	resnet-152_v2-tf	.264
26	resnet50-SIN_IN_IN	.264
27	resnet_SIN_IN_FT_IN	.264
28	densenet-201-keras	.263
29	resnet-50-pytorch	.257
30	resnet50-sup	.257
31	BiT-S-R152x4	.256
32	BiT-S-R152x2	.256
33	resnet-101_v2-tf	.251
34	inception_v3-tf	.250
35	densenet-169-keras	.249
36	BiT-S-R101x3	.247
37	resnet50	.243
38	BiT-S-R50x3	.241
39	BiT-S-R101x1	.241
40	resnet50-SIN_IN	.239
41	inception_v2-tf	.234
42	voneresnet-50-non_stochastic	.232
43	resnet-152_v1-tf	.231
44	resnet-101_v1-tf	.226
45	resnet-50_v2-tf	.225
46	BiT-S-R50x1	.224
47	regnet_y_400mf	.223
48	AdvProp_efficientnet-b8	.221
49	mobilenet_v2_1.4_224-tf	.221
50	AdvProp_efficientnet-b0	.220
51	resnet-34	.216
52	vonegrcnn_62e_nobn	.216
53	vonegrcnn_52e_full	.215
54	CORnet-S	.213
55	vonegrcnn_47e	.211
56	mobilenet_v2_1.0_224-tf	.209
57	mobilenet_v2_1.3_224-tf	.207
58	resnet-50_v1-tf	.205
59	inception_v1-tf	.191
60	vgg_19	.190
61	mobilenet_v2_0.75_224-tf	.187
62	nasnet_mobile-tf	.187
63	shufflenet_v2_x1_0	.182
64	resnet-18	.177
65	mobilenet_v2_1.0_192-tf	.176
66	resnet50-VITO-8deg-cc	.174
67	effnetb1_cutmixpatch_augmix_robust32_avge4e7_manylayers_324x288	.173
68	mobilenet_v2_0.75_192-tf	.165
69	resnet50-moclr8deg	.165
70	resnet50_robust_l2_eps1	.165
71	mobilenet_v1_1.0_224	.163
72	mobilenet_v2_1.0_160-tf	.162
73	resnet50_moco_v2	.159
74	custom_model_cv_18_dagger_408	.158
75	vgg-19-keras	.155
76	mobilenet_v2_0.5_224-tf	.154
77	mobilenet_v2_0.75_160-tf	.147
78	mobilenet_v1_0.75_224	.144
79	mobilenet_v1_1.0_192	.133
80	resnet50-SIN	.132
81	mobilenet_v2_1.0_128-tf	.132
82	mobilenet_v2_0.5_192-tf	.128
83	mobilenet_v1_1.0_160	.126
84	resnet50_finetune_cutmix_e3_robust_linf8255_e0_247x234	.126
85	resnet50_finetune_cutmix_AVGe2e3_robust_linf8255_e0_247x234	.125
86	mobilenet_v2_0.75_128	.122
87	mobilenet_v2_0.35_224	.120
88	mobilenet_v1_0.75_192	.118
89	mobilenet_v1_0.5_224	.115
90	mobilenet_v2_0.5_160	.111
91	mobilenet_v1_1.0_128	.107
92	mobilenet_v2_1.0_96	.106
93	mobilenet_v1_0.75_160	.105
94	mobilenet_v2_0.75_96	.100
95	mobilenet_v2_0.35_192	.097
96	0.5x_resnet-18_LC_w_sh_1_iter	.097
97	squeezenet1_1	.095
98	resnet50_l2_3_robust	.095
99	mobilenet_v1_0.75_128	.093
100	squeezenet1_0	.092
101	mobilenet_v2_0.35_160	.090
102	mobilenet_v1_0.5_192	.089
103	resnet-50-robust	.089
104	imagenet_l2_3_0	.089
105	resnet50_robust_l2_eps3	.089
106	mobilenet_v1_0.5_160	.083
107	mobilenet_v2_0.5_96	.075
108	alexnet_random_linf8_perturb	.072
109	mobilenet_v1_0.5_128	.071
110	resnet50_linf_8_robust	.070
111	alexnet	.069
112	alexnet-baseline	.069
113	alexnet	.069
114	alexnet	.069
115	alexnet	.069
116	alexnet_ks_torevert	.068
117	mobilenet_v2_0.35_128	.068
118	mobilenet_v1_0.25_224	.063
119	alexnet_early_checkpoint	.061
120	mobilenet_v2_0.35_96	.054
121	mobilenet_v1_0.25_192	.052
122	AlexNet_SIN	.040
123	mobilenet_v1_0.25_160	.040
124	mobilenet_v1_0.25_128	.038
125	alexnet_l2_3_robust	.035
126	imagenet_l2_10_0	.033
127	alexnet_robust_correct	.026
128	alexnet_linf_8_robust	.019
129	tf_efficientnetv2_s_in21ft1k_robust_linf12255_400x400	.009
130	barlow-twins-resnet50	.001
131	resnet-50_untrained	.000
132	resnet50-meshes-lt-100-original-pretrained	.000
133	resnet50-meshes-lt-100-original-scratch	.000
134	0.5x_resnet-18_LC_w_sh_100_iter
135	0.5x_resnet-18_LC_w_sh_10_iter
136	RN50
137	0.5x_resnet-18_LC
138	BiT-M-R50x3
139	BiT-M-R101x3
140	BiT-M-R152x4
141	BiT-M-R50x1
142	BiT-M-R152x2
143	BiT-M-R101x1
144	voneresnet-50-robust
145	resnet-18_untrained
146	fitvid_trained_on_physion
147	ViT_L_32
148	resnet18-supervised
149	FrankRobWobv0
150	resnet18-local_aggregation
151	resnet18-contrastive_multiview
152	ViT-B/32
153	resnet18-instance_recognition
154	resnet18-colorization
155	resnet18-deepcluster
156	resnet18-simclr
157	resnet18-relative_position
158	resnet-18-LC_w_sh_1_iter_conv_init_m
159	resnet18-contrastive_predictive
160	r3m_resnet18
161	resnet18-depth_prediction
162	resnet-18-LC_m
163	resnet-18-LC_w_sh_10_iter_conv_init_m
164	resnet18-autoencoder
165	resnet-18-LC_conv_init_m
166	r3m_resnet50
167	resnet-18_test_m
168	r3m_resnet34
169	resnet-18-LC_1st_conv_m
170	resnet-18-LC_w_sh_10_iter_m
171	r3m_resnet50_nocrop
172	CORnet-Z
173	resnet-18-LC_w_sh_100_iter_m
174	r3m_resnet34_nocrop
175	resnet-18-LC_w_sh_1_iter_m
176	RN50
177	resnet-18-LC_w_sh_100_iter_conv_init_m
178	0.5x_resnet-18
179	resnet-18-LC_d_w_sh_1x1_conv_init_m
180	CORnetZ_CIFAR10
181	vggface
182	my-model
183	omnivore_swinS
184	omnivore_swinT
185	omnivore_swinB
186	resnet50
187	openclip
188	fulltest_microblockvf_nobottleneck_freeat_nopretrain_eps2_m4
189	antialiased-rnext101_32x8d
190	Res2Net50_26w_4s
191	yudixie_resnet50_imagenet1kpret_0_240312
192	resnet18_imagenet21kP
193	ConvLSTM
194	PredRNN
195	TAU
196	MIM
197	resnext101_32x32d_wsl

Benchmark bibtex

@inproceedings{DBLP:conf/nips/BarbuMALWGTK19,
                                                    author    = {Andrei Barbu and
                                                                David Mayo and
                                                                Julian Alverio and
                                                                William Luo and
                                                                Christopher Wang and
                                                                Dan Gutfreund and
                                                                Josh Tenenbaum and
                                                                Boris Katz},
                                                    title     = {ObjectNet: {A} large-scale bias-controlled dataset for pushing the
                                                                limits of object recognition models},
                                                    booktitle = {NeurIPS 2019},
                                                    pages     = {9448--9458},
                                                    year      = {2019},
                                                    url       = {https://proceedings.neurips.cc/paper/2019/hash/97af07a14cacba681feacf3012730892-Abstract.html},
                                                    }

Ceiling

1.00.

Note that scores are relative to this ceiling.

Data: ObjectNet

Metric: top1