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("Hebart2023_fmri.V2-ridgecv")
score = benchmark(my_model)

Benchmark API

Code examples

Model scores

Score Legend Min Alignment Max Alignment

Rank	Model	Score
1	deit3_base_patch16_224_fb_in22k_ft_in1k_gram_aligned_dinov3	.661
2	deit3_base_patch16_224_fb_in22k_ft_in1k	.661
3	cvt_cvt-13-384-in1k_4_LucyV4	.644
4	swin_small_patch4_window7_224:ms_in22k_ft_in1k	.644
5	deit_small_imagenet_full_seed-0	.643
6	vit_tiny_r_s16_p8_384:augreg_in21k_ft_in1k	.640
7	cvt_cvt-w24-384-in22k_finetuned-in1k_4	.636
8	vit_large_clip336_openai	.633
9	dinov2_vitl14_lc_res336	.631
10	dinov2_vitl14_reg_lc_featread	.631
11	dinov2_vitl14_reg_lc	.631
12	deit3_base_patch16_224_fb_in1k_gram_aligned_dinov3	.630
13	deit3_base_patch16_224_fb_in1k	.630
14	convnext_large:fb_in22k_ft_in1k	.630
15	resnet50_st_wzc_seed7_2_323	.628
16	resnet50_st_wzc_seed6_2_323	.627
17	resnet50-SIN_IN_IN	.626
18	resnet_SIN_IN_FT_IN	.626
19	dinov2_vitl14_lc	.625
20	resnet50_st_wzc_seed3_2_323	.625
21	dinov2_vitg14_lc	.622
22	dinov2_vitg14_reg_lc	.621
23	convnext_large_in22k_supervised	.618
24	antialiased-r50	.618
25	cvt_cvt-21-224-in1k_4_LucyV4	.614
26	Res2Net50_26w_4s	.613
27	resnext101_32x16d_wsl	.612
28	deit_base_imagenet_full_seed-0	.611
29	resnet152_imagenet_full	.610
30	resnet50_ewconplus_wzc_seed5_2__323	.610
31	voneresnet-50-non_stochastic	.609
32	resnext101_32x8d_wsl	.609
33	focalnet_tiny_lrf_in1k	.608
34	eva02_large_in22k	.607
35	resnet50_ewconplus_wzc_seed6_2_323	.607
36	resnet_50_v1	.607
37	resnet50-sup	.607
38	resnet_50_v1_spiking	.607
39	resnet50_imagenet_full	.606
40	cvt_cvt-13-384-in22k_finetuned-in1k_4_LucyV4	.605
41	resnet50-SIN_IN	.605
42	resnet50-vitoimagevidnet8	.603
43	resnet50-VITO-8deg-cc	.603
44	deit_large_imagenet_full_seed-0	.603
45	resnet50_ewconplus_wzc_seed2_2__323	.603
46	cvt_cvt-21-384-in22k_finetuned-in1k_4_LucyV4	.603
47	resnet50_robust_l2_eps1	.602
48	resnet50-moclr8deg	.600
49	vit_relpos_base_patch16_clsgap_224:sw_in1k	.599
50	vit_large_patch14_clip_336:openai_ft_in12k_in1k	.599
51	cvt_cvt-21-384-in1k_4_LucyV4	.598
52	resnext101_32x32d_wsl	.597
53	resnet_50_v2	.595
54	resnet50_ewconplus_wzc	.596
55	voneresnet-50	.593
56	densenet-121	.592
57	convnext_large_imagenet_full_seed-0	.592
58	convnext_xxlarge:clip_laion2b_soup_ft_in1k	.592
59	custom_model_cv_18_dagger_408	.591
60	resnext101_32x48d_wsl	.590
61	convnext_tiny_sup	.590
62	shufflenet_v2_x1_0	.589
63	vit_large_patch14_clip_336:laion2b_ft_in1k	.588
64	vit_base_patch16_clip_224:openai_ft_in12k_in1k	.587
65	resnet50_tutorial	.586
66	AdvProp_efficientnet-b4	.584
67	deepeisnn_ei_imagenet_resnet18	.584
68	AdvProp_efficientnet-b8	.583
69	resnet50-SIN	.582
70	convnext_xlarge:fb_in22k_ft_in1k	.581
71	vit_large_patch14_clip_224:openai_ft_in1k	.579
72	vit_large_patch14_clip_224:openai_ft_in12k_in1k	.578
73	yudixie_wbh_object_class_scaling_0415_imagenet	.577
74	vit_huge_patch14_clip_224:laion2b_ft_in12k_in1k	.577
75	AdvProp_efficientnet-b6	.577
76	cvt_cvt-13-224-in1k_4_LucyV4	.576
77	convnext_tiny_imagenet_full_seed-0	.572
78	efficientnet_b2_imagenet_full	.572
79	AdvProp_efficientnet-b7	.571
80	convnext_xxlarge_clip	.570
81	convnext_femto_ols:d1_in1k	.569
82	vit_large_patch14_clip_224:laion2b_ft_in1k	.568
83	vit_large_patch14_clip_224:laion2b_ft_in12k_in1k	.565
84	grcnn	.563
85	mobilenet_v2_1_4_224	.562
86	convnext_large_clip_laion2b_finelayers	.561
87	convnext_tiny:in12k_ft_in1k	.554
88	convnext_base:clip_laiona_augreg_ft_in1k_384	.550
89	effnetb1_cutmix_augmix_sam_e1_5avg_424x377	.537
90	convnext_large_mlp:clip_laion2b_augreg_ft_in1k_384	.535
91	convnext_small_imagenet_full_seed-0	.534
92	ReAlnet01_cornet	.530
93	alexnet	.527
94	convnext_base_imagenet_full_seed-0	.526
95	vit_base_patch16_clip_224:openai_ft_in1k	.525
96	ReAlnet03_cornet	.524
97	convnext_base_fb_in22k_ft_in1k_gram_aligned_dinov3	.523
98	convnext_base_fb_in1k	.520
99	ReAlnet04_cornet	.520
100	vit_relpos_base_patch32_plus_rpn_256:sw_in1k	.517
101	effnetb1_272x240	.519
102	AlexNet_SIN	.512
103	AT_efficientnet-b2	.509
104	AdvProp_efficientnet-b2	.509
105	effnetb1_cutmixpatch_SAM_robust32_avge6e8e9e10_manylayers_324x288	.509
106	efficientnet_b0	.505
107	resnet-50-robust	.497
108	AlexNet_SIN_fov12	.496
109	AlexNet_SIN_fov16	.482
110	resnet18_fair	.480
111	effnetb1_cutmixpatch_augmix_robust32_avge4e7_manylayers_324x288	.478
112	VOneCORnet-S	.477
113	deepeisnn_ei_cifar10_vgg19	.465
114	deepeisnn_ei_cifar10_resnet18	.435
115	alexnet_training_seed_06	.409
116	alexnet_training_seed_07	.390
117	alexnet_training_seed_01	.383
118	yudixie_wbh_object_class_scaling_0415_run_2	.381
119	alexnet_training_seed_03	.379
120	alexnet_training_seed_10	.375
121	alexnet_training_seed_02	.369
122	ReAlnet01_cornet_fov4	.370
123	yudixie_wbh_object_class_scaling_0415_run_0	.350
124	yudixie_wbh_object_class_scaling_0415_run_3	.350
125	alexnet_training_seed_09	.350
126	predictive_coding_pc	.351
127	pr1_stl10_disentangled_alpha04	.348
128	yudixie_wbh_object_class_scaling_0415_run_1	.337
129	ViT_L_32_imagenet1k	.308
130	convnext_base_fb_in22k_ft_in1k	.225
131	compact_resnet50_robust_V4	.183
132	convnext_base_fb_in1k_gram_aligned_dinov3	.182
133	vit_huge_clip336_laion
134	pnasnet_large_halfPrecision
135	hmax

Benchmark bibtex

None

Ceiling

0.48.

Note that scores are relative to this ceiling.

Data: Hebart2023_fmri.V2

Metric: ridgecv