Sample stimuli

How to use

from brainscore_vision import load_benchmark
benchmark = load_benchmark("Maniquet2024-tasks_consistency")
score = benchmark(my_model)

Benchmark API

Code examples

Model scores

Score Legend Min Alignment Max Alignment

Rank	Model	Score
1	vit_base_patch16_clip_224:openai_ft_in1k	.750
2	efficientnet_b2_imagenet_full	.748
3	effnetb1_272x240	.746
4	cvt_cvt-21-384-in1k_4_LucyV4	.746
5	vit_tiny_r_s16_p8_384:augreg_in21k_ft_in1k	.745
6	convnext_femto_ols:d1_in1k	.741
7	cvt_cvt-21-384-in22k_finetuned-in1k_4_LucyV4	.739
8	vit_large_patch14_clip_224:laion2b_ft_in1k	.738
9	efficientnet_b1_imagenet_full	.737
10	convnext_large_mlp:clip_laion2b_augreg_ft_in1k_384	.735
11	resnet18_imagenet21kP	.734
12	effnetb1_cutmix_augmix_sam_e1_5avg_424x377	.734
13	cvt_cvt-13-384-in1k_4_LucyV4	.732
14	resnet_152_v2	.729
15	efficientnet_b0_imagenet_full	.727
16	vgg_19	.727
17	mobilenet_v2_1_0_224	.724
18	resnet-152_v2_pytorch	.718
19	resnet50-vicregl0p9	.717
20	densenet-201	.716
21	cvt_cvt-21-224-in1k_4_LucyV4	.716
22	antialiased-r50	.715
23	cvt_cvt-13-224-in1k_4_LucyV4	.715
24	nasnet_large	.713
25	mobilenet_v2_1-4_224_pytorch	.712
26	mobilenet_v2_1_4_224	.712
27	resnext101_32x32d_wsl	.710
28	resnet_50_v2	.707
29	resnet_101_v1	.705
30	resnet101_ecoset_full	.705
31	resnet18_ecoset_full	.703
32	deit_base_imagenet_full_seed-0	.701
33	resnet152_imagenet_full	.697
34	Res2Net50_26w_4s	.692
35	resnext101_32x16d_wsl	.692
36	yudixie_resnet18_imagenet1kpret_0_240719	.690
37	vgg_16	.689
38	inception_v3	.688
39	mobilenet_v2_0_75_160	.688
40	resnet50_ecoset_full	.686
41	resnet50-vicregl0p75	.685
42	convnext_tiny_sup	.684
43	resnet_101_v2	.684
44	efficientnet_b0	.683
45	deit_small_imagenet_full_seed-0	.681
46	tv_efficientnet-b1	.681
47	resnet34_ecoset_full	.680
48	vonegrcnn_52e_full	.679
49	ReAlnet02_cornet	.678
50	ReAlnet02	.678
51	vonegrcnn_62e_nobn	.676
52	resnet50-sup	.676
53	resnet_50_v1	.676
54	regnet_y_400mf	.676
55	resnet-18_test_m	.674
56	deit_large_imagenet_full_seed-0	.674
57	grcnn	.672
58	yudixie_resnet50_imagenet1kpret_0_240908	.672
59	resnet50_eMMCR_Vanilla	.671
60	resnet50-SIN_IN_IN	.669
61	inception_v1	.669
62	convnext_small_imagenet_full_seed-0	.668
63	vonegrcnn_47e	.668
64	resnet50_tutorial	.667
65	nasnet_mobile	.667
66	ReAlnet06_cornet	.667
67	ReAlnet06	.667
68	resnet18_imagenet_full	.667
69	resnext101_32x8d_wsl	.667
70	convnext_tiny:in12k_ft_in1k	.666
71	cvt_cvt-13-384-in22k_finetuned-in1k_4_LucyV4	.666
72	convnext_small_imagenet_100_seed-0	.666
73	convnext_tiny_imagenet_full_seed-0	.665
74	xception	.663
75	vit_relpos_base_patch16_clsgap_224:sw_in1k	.661
76	resnet34_imagenet_full	.660
77	BiT-S-R50x1	.660
78	resnet50_imagenet_full	.659
79	inception_v3_pytorch	.659
80	vit_large_patch14_clip_224:openai_ft_in12k_in1k	.658
81	resnet_SIN_IN_FT_IN	.657
82	ReAlnet05_cornet	.656
83	ReAlnet05	.656
84	convnext_xlarge:fb_in22k_ft_in1k	.656
85	shufflenet_v2_x1_0	.656
86	ReAlnet03_cornet	.656
87	ReAlnet03	.656
88	resnet101_imagenet_full	.654
89	ReAlnet04_cornet	.653
90	ReAlnet04	.653
91	resnet152_ecoset_full	.651
92	convnext_large_imagenet_full_seed-0	.650
93	ReAlnet09_cornet	.649
94	ReAlnet09	.649
95	vit_base_patch16_clip_224:openai_ft_in12k_in1k	.649
96	resnet50-moclr8deg	.648
97	ReAlnet01_cornet	.648
98	ReAlnet_incorrect	.648
99	ReAlnet01	.648
100	inception_v4	.648
101	ReAlnet07_cornet	.647
102	ReAlnet07	.647
103	ReAlnet08	.647
104	ReAlnet08_cornet	.647
105	resnet_152_v1	.646
106	ReAlnet10	.646
107	ReAlnet10_cornet	.646
108	resnet50-vicreg	.646
109	resnext101_32x48d_wsl	.644
110	swin_small_patch4_window7_224:ms_in22k_ft_in1k	.644
111	cvt_cvt-w24-384-in22k_finetuned-in1k_4	.642
112	resnet50-vitoimagevidnet8	.640
113	resnet50_imagenet_100_seed-0	.638
114	antialias-resnet152	.638
115	VOneCORnet-S	.638
116	convnext_base_imagenet_full_seed-0	.635
117	mobilenet_v2_1_0_160	.632
118	voneresnet-50-non_stochastic	.625
119	grcnn_109	.625
120	alexnet_ks_torevert	.624
121	effnetb1_cutmixpatch_augmix_robust32_avge4e7_manylayers_324x288	.617
122	BiT-S-R101x1	.617
123	effnetb1_cutmixpatch_SAM_robust32_avge6e8e9e10_manylayers_324x288	.616
124	vit_large_patch14_clip_224:laion2b_ft_in12k_in1k	.616
125	resnet50-SIN_IN	.610
126	voneresnet-50	.609
127	deit_small_imagenet_100_seed-0	.606
128	ViT_L_32_imagenet1k	.604
129	resnet50-barlow	.596
130	vit_relpos_base_patch32_plus_rpn_256:sw_in1k	.576
131	mobilenet_v2_1_0_128	.568
132	AT_efficientnet-b2	.565
133	AdvProp_efficientnet-b2	.565
134	focalnet_tiny_lrf_in1k	.550
135	yudixie_resnet50_object_class_0_240908	.550
136	yudixie_resnet50_distance_translation_0_240908	.549
137	yudixie_resnet50_cat_obj_class_all_latents_0_240908	.545
138	yudixie_resnet18_distance_rotation_0_240719	.541
139	resnet50_robust_l2_eps1	.541
140	yudixie_resnet18_category_class_0_240719	.535
141	blt_vs	.534
142	yudixie_resnet18_object_class_0_240719	.531
143	yudixie_resnet18_translation_rotation_0_240719	.531
144	yudixie_resnet18_cat_obj_class_all_latents_0_240719	.525
145	yudixie_resnet50_distance_translation_rotation_0_240908	.521
146	yudixie_resnet18_distance_translation_0_240719	.521
147	yudixie_resnet50_category_class_0_240908	.520
148	efficientnet-b7	.507
149	yudixie_resnet50_rotation_reg_0_240908	.502
150	AlexNet_SIN	.499
151	yudixie_resnet50_translation_reg_0_240908	.498
152	yudixie_resnet50_distance_rotation_0_240908	.498
153	artResNet18_1	.494
154	yudixie_resnet18_translation_reg_0_240719	.484
155	yudixie_resnet18_rotation_reg_0_240719	.484
156	convnext_small_imagenet_10_seed-0	.484
157	yudixie_resnet18_distance_translation_rotation_0_240719	.482
158	resnet50-SIN	.479
159	resnet50_imagenet_10_seed-0	.478
160	yudixie_resnet50_translation_rotation_0_240908	.476
161	yudixie_resnet18_distance_reg_0_240719	.470
162	pnasnet_large_pytorch	.470
163	yudixie_resnet18_random_0_240719	.462
164	AdvProp_efficientnet-b4	.461
165	yudixie_resnet50_distance_reg_0_240908	.450
166	resnet-50-robust	.437
167	imagenet_l2_3_0	.395
168	resnet50_robust_l2_eps3	.395
169	AdvProp_efficientnet-b6	.383
170	alexnet_training_seed_10	.380
171	alexnet_training_seed_06	.379
172	alexnet_training_seed_02	.367
173	alexnet_training_seed_01	.358
174	alexnet_training_seed_09	.349
175	alexnet_training_seed_03	.349
176	alexnet_training_seed_07	.344
177	alexnet_training_seed_05	.343
178	yudixie_resnet50_random_0_240908	.341
179	alexnet_training_seed_04	.326
180	alexnet_training_seed_08	.325
181	resnet50_imagenet_1_seed-0	.323
182	deit_small_imagenet_1_seed-0	.296
183	convnext_small_imagenet_1_seed-0	.266
184	deit_small_imagenet_10_seed-0	.210
185	omnivore_swinS	X
186	densenet_201_pytorch	X
187	BiT-S-R152x4	X
188	BiT-S-R101x3	X
189	BiT-S-R152x2	X
190	MIM	X
191	TAU	X
192	SimVP	X
193	PredRNN	X
194	pnasnet_large	X
195	antialiased-rnext101_32x8d	X
196	custom_model_cv_18_dagger_408	X
197	densenet-169	X
198	densenet-121	X
199	convnext_base:clip_laiona_augreg_ft_in1k_384	X
200	hmax	X
201	vit_large_patch14_clip_224:openai_ft_in1k	X
202	convnext_large:fb_in22k_ft_in1k	X
203	vit_large_patch14_clip_336:openai_ft_in12k_in1k	X
204	vit_large_patch14_clip_336:laion2b_ft_in1k	X
205	mobilenet_v2_1_3_224	X
206	mobilenet_v2_0_5_192	X
207	omnivore_swinB	X
208	voneresnet-50-robust	X
209	mobilenet_v2_0_75_192	X
210	mobilenet_v2_0_5_224	X
211	AdvProp_efficientnet-b7	X
212	vit_huge_patch14_clip_336:laion2b_ft_in12k_in1k	X
213	BiT-S-R50x3	X
214	vit_huge_patch14_clip_224:laion2b_ft_in12k_in1k	X
215	mobilenet_v2_0_75_224	X
216	resnet50-VITO-8deg-cc	X
217	mobilenet_v2_1_0_192	X
218	ConvLSTM	X
219	AdvProp_efficientnet-b8	X
220	convnext_xxlarge:clip_laion2b_soup_ft_in1k	X

Benchmark bibtex

@article {Maniquet2024.04.02.587669,
	author = {Maniquet, Tim and de Beeck, Hans Op and Costantino, Andrea Ivan},
	title = {Recurrent issues with deep neural network models of visual recognition},
	elocation-id = {2024.04.02.587669},
	year = {2024},
	doi = {10.1101/2024.04.02.587669},
	publisher = {Cold Spring Harbor Laboratory},
	URL = {https://www.biorxiv.org/content/early/2024/04/10/2024.04.02.587669},
	eprint = {https://www.biorxiv.org/content/early/2024/04/10/2024.04.02.587669.full.pdf},
	journal = {bioRxiv}
}

Ceiling

Not available

Data: Maniquet2024

Metric: tasks_consistency