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

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