Sample stimuli

How to use

from brainscore_vision import load_benchmark
benchmark = load_benchmark("Marques2020_Cavanaugh2002-surround_diameter")
score = benchmark(my_model)

Benchmark API

Code examples

Model scores

Score Legend Min Alignment Max Alignment

Rank	Model	Score
1	resnet50-VITO-8deg-cc	.983
2	mobilenet_v2_0.75_224-tf	.968
3	mobilenet_v2_0_75_224	.964
4	mobilenet_v2_1.0_192-tf	.953
5	resnet-101_v2-tf	.948
6	convnext_large:fb_in22k_ft_in1k	.940
7	nasnet_mobile	.938
8	mobilenet_v2_0_5_192	.932
9	mobilenet_v2_0_5_224	.926
10	mobilenet_v2_0_75_160	.922
11	resnet-50_v2-tf	.919
12	convnext_femto_ols:d1_in1k	.899
13	imagenet_l2_10_0	.887
14	inception_resnet_v2	.871
15	convnext_xlarge:fb_in22k_ft_in1k	.870
16	omnivore_swinB	.863
17	convnext_large_mlp:clip_laion2b_augreg_ft_in1k_384	.863
18	resnet-101_v1-tf	.857
19	cvt_cvt-21-384-in1k_4_LucyV4	.848
20	resnet50-SIN	.848
21	convnext_tiny_sup	.847
22	vit_tiny_r_s16_p8_384:augreg_in21k_ft_in1k	.844
23	efficientnet_b0_imagenet_full	.839
24	imagenet_l2_3_0	.838
25	cvt_cvt-21-224-in1k_4_LucyV4	.834
26	convnext_small_imagenet_full_seed-0	.832
27	resnet-152_v1-tf	.829
28	nasnet_mobile-tf	.828
29	omnivore_swinS	.827
30	mobilenet_v2_0.5_224-tf	.824
31	resnet50-vicreg	.815
32	alexnet_linf_8_robust	.810
33	mobilenet_v2_1_0_224	.808
34	mobilenet_v2_1-4_224_pytorch	.807
35	convnext_base:clip_laiona_augreg_ft_in1k_384	.807
36	efficientnet_b1_imagenet_full	.805
37	convnext_small_imagenet_100_seed-0	.795
38	cvt_cvt-13-224-in1k_4_LucyV4	.794
39	BiT-M-R152x4	.794
40	nasnet_large-tf	.789
41	resnet50_robust_l2_eps1	.789
42	inception_v3	.785
43	0.5x_resnet-18_LC_w_sh_100_iter	.782
44	effnetb1_272x240	.781
45	FrankRobWobv0	.780
46	alexnet_robust_correct	.779
47	resnet50-barlow	.775
48	vit_relpos_base_patch32_plus_rpn_256:sw_in1k	.770
49	texture_shape_resnet50_trained_on_SIN	.768
50	resnet50_l2_3_robust	.767
51	resnet50-SIN_IN_IN	.765
52	resnet_SIN_IN_FT_IN	.765
53	RN50	.764
54	resnet50_moco_v2	.762
55	mobilenet_v2_1_3_224	.756
56	convnext_base_imagenet_full_seed-0	.756
57	resnet50-moclr8deg	.754
58	BiT-S-R50x3	.754
59	konkle_alexnetgn_ipcl_ref12_supervised_ipcl_aug	.753
60	resnet-50-robust	.752
61	CLIP_resnet50_float32	.751
62	CLIP_resnet50	.752
63	inception_v3-tf	.752
64	deit_small_imagenet_100_seed-0	.751
65	cvt_cvt-13-384-in22k_finetuned-in1k_4_LucyV4	.749
66	yudixie_resnet18_imagenet1kpret_0_240719	.749
67	yudixie_resnet50_imagenet1kpret_0_240908	.747
68	resnet_50_v1	.747
69	resnet50-sup	.748
70	resnet50_tutorial	.747
71	resnet50_linf_8_robust	.746
72	alexnet_training_seed_09	.744
73	convnext_tiny:in12k_ft_in1k	.742
74	antialiased-rnext101_32x8d	.742
75	fulltest_microblockvf_nobottleneck_freeat_nopretrain_eps2_m4	.743
76	alexnet_training_seed_10	.742
77	SWSL_resnext101_32x8d	.742
78	resnet50_byol	.740
79	mobilevit_small	.740
80	resnet50	.738
81	resnet50_linf_4_robust	.737
82	alexnet_l2_3_robust	.736
83	resnet50_robust_l2_eps3	.736
84	resnet-18-LC_w_sh_100_iter_conv_init_m	.736
85	mobilenet_v2_0.35_224	.736
86	mobilenet_v2_1_0_160	.738
87	texture_shape_alexnet_trained_on_SIN	.736
88	resnet50-vicregl0p75	.735
89	SWSL_resnet50	.736
90	resnet50_FractalDB	.735
91	pnasnet_large-tf	.735
92	alexnet_training_seed_01	.734
93	resnet-50_v1-tf	.733
94	convnext_large_imagenet_full_seed-0	.732
95	resnet18_imagenet21kP	.729
96	efficientnet_b0	.728
97	BiT-S-R152x4	.727
98	alexnet_training_seed_07	.726
99	resnet50_random_l2_perturb	.725
100	alexnet_training_seed_02	.723
101	yudixie_resnet18_object_class_0_240719	.719
102	0.5x_resnet-18_LC	.713
103	alexnet_random_l2_3_perturb	.710
104	alexnet	.710
105	blt_vs	.691
106	BiT-S-R152x2	.688
107	CLIP-RN50	.689
108	konkle_alexnetgn_ipcl_ref01_primary_model	.686
109	inception_v2-tf	.681
110	BiT-S-R101x3	.680
111	resnet50_random_linf8_perturb	.679
112	convnext_tiny_imagenet_full_seed-0	.677
113	alexnet_training_seed_04	.674
114	inception_v4-tf	.669
115	ViT_L_32_imagenet1k	.664
116	cvt_cvt-13-384-in1k_4_LucyV4	.664
117	convnext_xxlarge:clip_laion2b_soup_ft_in1k	.664
118	mobilenet_v2_0.75_96	.658
119	mobilenet_v2_0.5_160	.658
120	mobilenet_v1_0.25_128	.658
121	effnetb1_cutmix_augmix_sam_e1_5avg_424x377	.652
122	efficientnet_b2_imagenet_full	.635
123	cvt_cvt-w24-384-in22k_finetuned-in1k_4	.632
124	resnet50_simclr	.627
125	effnetb1_cutmixrespatch_SAM_robust32_e10_manylayers_324x288	.628
126	effnetb1_cutmixpatch_SAM_robust32_avge6e8e9e10_manylayers_324x288	.626
127	effnetb1_cutmixpatch_augmix_robust32_e5_324x288	.624
128	effnetb1_cutmixpatch_augmix_robust32_avge4e7_manylayers_324x288	.624
129	effnetb1_cutmixrespatch_SAM_robust32_e2_manylayers_324x288	.622
130	BiT-S-R101x1	.621
131	omnivore_swinT	.620
132	resnet-18-LC_conv_init_m	.619
133	vit_large_patch14_clip_224:laion2b_ft_in12k_in1k	.613
134	densenet-121-keras	.611
135	r3m_resnet34	.610
136	resnet50_finetune_cutmix_AVGe2e3_robust_linf8255_e0_247x234	.609
137	resnet50_finetune_cutmix_e3_robust_linf8255_e0_247x234	.609
138	effnetv2m_custom384	.603
139	deit_large_imagenet_full_seed-0	.601
140	antialiased-r50	.601
141	resnet-152_v2-tf	.598
142	cv_18_dagger_408_pretrained	.597
143	vgg-16-keras	.592
144	CORnetZ_CIFAR10	.589
145	BiT-S-R50x1	.588
146	r3m_resnet18	.587
147	BiT-M-R101x3	.584
148	my-model	.582
149	cvt_cvt-21-384-in22k_finetuned-in1k_4_LucyV4	.580
150	densenet-169-keras	.578
151	deit_base_imagenet_full_seed-0	.567
152	densenet-169	.557
153	resnet-18-LC_w_sh_1_iter_conv_init_m	.556
154	vgg_16	.555
155	vggface	.555
156	swin_small_patch4_window7_224:ms_in22k_ft_in1k	.554
157	tf_efficientnetv2_s_in21ft1k_robust_linf12255_400x400	.551
158	convnext_small_imagenet_10_seed-0	.550
159	r3m_resnet50_nocrop	.545
160	yudixie_resnet18_random_0_240719	.543
161	BiT-M-R152x2	.541
162	BiT-M-R50x3	.537
163	vgg-19-keras	.534
164	vit_large_patch14_clip_224:openai_ft_in1k	.532
165	densenet-121	.531
166	vit_large_patch14_clip_224:laion2b_ft_in1k	.525
167	AdvProp_efficientnet-b2	.519
168	densenet-201-keras	.518
169	AT_efficientnet-b2	.517
170	vit_relpos_base_patch16_clsgap_224:sw_in1k	.516
171	BiT-M-R50x1	.512
172	focalnet_tiny_lrf_in1k	.509
173	resnet-18-LC_w_sh_10_iter_m	.507
174	resnet-152_v2_pytorch	.506
175	mobilenet_v2_0.75_192-tf	.505
176	shufflenet_v2_x1_0	.504
177	vit_huge_patch14_clip_224:laion2b_ft_in12k_in1k	.503
178	AdvProp_efficientnet-b4	.502
179	antialias-resnet152	.498
180	grcnn_robust_v1	.497
181	vit_large_patch14_clip_224:openai_ft_in12k_in1k	.494
182	BiT-M-R101x1	.494
183	resnet-18-LC_w_sh_100_iter_m	.493
184	resnet-18-LC_w_sh_10_iter_conv_init_m	.491
185	custom_model_cv_18_dagger_408	.489
186	CLIP_ViT-B_32	.481
187	r3m_resnet50	.480
188	AdvProp_efficientnet-b8	.476
189	resnet-18-LC_1st_conv_m	.470
190	grcnn	.467
191	xception	.466
192	grcnn_v2_text_noise	.462
193	vit_huge_patch14_clip_336:laion2b_ft_in12k_in1k	.451
194	AdvProp_efficientnet-b7	.448
195	yudixie_resnet50_random_0_240908	.447
196	mobilenet_v2_1_0_192	.442
197	alexnet_training_seed_03	.440
198	vit_large_patch14_clip_336:laion2b_ft_in1k	.440
199	efficientnet-b7	.433
200	grcnn_109	.427
201	yudixie_resnet18_translation_reg_0_240719	.426
202	mobilenet_v2_1.3_224-tf	.426
203	yudixie_resnet18_cat_obj_class_all_latents_0_240719	.425
204	AdvProp_efficientnet-b0	.423
205	mobilenet_v2_1_4_224	.423
206	resnet-18-LC_d_w_sh_1x1_conv_init_m	.418
207	yudixie_resnet18_distance_translation_0_240719	.417
208	yudixie_resnet18_translation_rotation_0_240719	.417
209	yudixie_resnet18_distance_rotation_0_240719	.414
210	densenet-201	.414
211	resnet-18-LC_m	.411
212	yudixie_resnet18_distance_translation_rotation_0_240719	.409
213	mobilenet_v2_0_75_192	.408
214	vonealexnet_gaussian_noise_std4_fixed	.404
215	yudixie_resnet18_distance_reg_0_240719	.399
216	yudixie_resnet18_category_class_0_240719	.397
217	yudixie_resnet50_rotation_reg_0_240908	.396
218	yudixie_resnet50_distance_reg_0_240908	.394
219	yudixie_resnet50_translation_rotation_0_240908	.395
220	yudixie_resnet50_translation_reg_0_240908	.394
221	yudixie_resnet50_distance_rotation_0_240908	.394
222	yudixie_resnet50_distance_translation_rotation_0_240908	.392
223	yudixie_resnet50_distance_translation_0_240908	.392
224	resnet101	.390
225	resnet-50x2_untrained	.389
226	hmax	.390
227	mobilenet_v2_1.0_224-tf	.389
228	mobilenet_v2_1.4_224-tf	.389
229	mobilenet_v2_1.0_128-tf	.389
230	mobilenet_v2_1.0_96	.389
231	resnet50-vitoimagevidnet8	.390
232	0.5x_resnet-18_LC_w_sh_10_iter	.390
233	0.5x_resnet-18_LC_w_sh_1_iter	.390
234	barlow-twins-resnet50	.389
235	resnet-18_test_m	.390
236	resnet-18-LC_w_sh_1_iter_m	.390
237	resnet-18_untrained	.390
238	resnet-50_untrained	.389
239	yudixie_resnet50_cat_obj_class_all_latents_0_240908	.390
240	yudixie_resnet50_category_class_0_240908	.389
241	yudixie_resnet18_rotation_reg_0_240719	.390
242	yudixie_resnet50_object_class_0_240908	.389
243	deit_small_imagenet_10_seed-0	.390
244	artResNet18_1	.389
245	inception_v1	.389
246	resnet_101_v1	.390
247	0.5x_resnet-18	.389
248	resnet50-simclr-vissl	.390
249	mobilenet_v2_0.5_192-tf	.389
250	resnet50-vicregl0p9	.389
251	ReAlnet02	.388
252	ReAlnet06	.389
253	ReAlnet10_cornet	.389
254	CORnet-S	.389
255	ReAlnet_incorrect	.389
256	resnet34_imagenet_full	.389
257	ReAlnet02_cornet	.389
258	ReAlnet08	.389
259	ReAlnet06_cornet	.388
260	ReAlnet01_cornet	.389
261	ReAlnet03_cornet	.389
262	resnet50-SIN_IN	.388
263	resnet18_ecoset_full	.388
264	ReAlnet10	.389
265	ReAlnet03	.388
266	ReAlnet01	.388
267	ReAlnet08_cornet	.388
268	ReAlnet07_cornet	.388
269	ReAlnet07	.387
270	ReAlnet04_cornet	.387
271	ReAlnet09	.387
272	ReAlnet09_cornet	.387
273	ReAlnet04	.387
274	alexnet	.386
275	ReAlnet05	.387
276	ReAlnet05_cornet	.386
277	alexnet	.386
278	alexnet	.385
279	alexnet	.385
280	alexnet	.385
281	alexnet	.385
282	resnet34_ecoset_full	.385
283	alexnet	.386
284	alexnet-baseline	.385
285	voneresnet-50-non_stochastic	.384
286	resnet18_imagenet_full	.384
287	vit_large_patch14_clip_336:openai_ft_in12k_in1k	.379
288	v1-pyr-nodown	.374
289	Res2Net50_26w_4s	.369
290	cornet_s	.368
291	resnet-18-LC_untrained	.361
292	vonegrcnn_47e	.358
293	vision_transformer_vit_large_patch16_224	.355
294	VOneCORnet-S	.344
295	vit_base_patch16_clip_224:openai_ft_in12k_in1k	.336
296	vonegrcnn_52e_full	.334
297	voneresnet-50	.329
298	resnext101_32x8d_wsl	.328
299	voneresnet-50-robust	.319
300	vonegrcnn_62e_nobn	.314
301	vonegrcnn_47e	.314
302	convnext_small_imagenet_1_seed-0	.312
303	alexnet_training_seed_05	.311
304	alexnet_training_seed_08	.305
305	vit_base_patch16_clip_224:openai_ft_in1k	.303
306	AdvProp_efficientnet-b6	.297
307	alexnet_training_seed_06	.281
308	resnet_50_v2	.280
309	resnet50_imagenet_1_seed-0	.269
310	alexnet_reduced_aliasing_early_checkpoint	.265
311	resnet50_imagenet_10_seed-0	.257
312	ViT-B/32	.255
313	resnet101_imagenet_full	.245
314	resnet50_ecoset_full	.222
315	resnet101_ecoset_full	.217
316	alexnet_random_linf8_perturb	.198
317	resnet152_ecoset_full	.197
318	alexnet_early_checkpoint	.190
319	vgg_19	.188
320	resnet152_imagenet_full	.177
321	CORnet-Z	.156
322	inception_v1-tf	.156
323	resnet50-meshes-lt-100-original-pretrained	.156
324	mobilenet_v2_1.0_160-tf	.156
325	AlexNet_SIN	.156
326	resnet50_imagenet_full	.156
327	resnet50_imagenet_100_seed-0	.156
328	deit_small_imagenet_full_seed-0	.028
329	CORnetZ_CIFAR10_bs32_20_04	.000
330	resnet50_pretrained_with_retinal_waves	.000
331	mobilenet_v2_0.5_128	.000
332	mobilenet_v2_0.75_128	.000
333	mobilenet_v2_0.75_160-tf	.000
334	resnet50-cifar	.000
335	my-model	.000
336	resnet50	.000
337	resnet50-meshes-lt-100-original-scratch	.000
338	mobilenet_v2_1_0_128	.000
339	pnasnet_large_pytorch	X
340	inception_v4	X
341	inception_v3_pytorch	X
342	pnasnet_large	X
343	fitvid_trained_on_physion	X
344	resnext101_32x32d_wsl	X
345	vonegrcnn_52e_full	X
346	resnext101_32x48d_wsl	X
347	openclip	X
348	yudixie_resnet50_imagenet1kpret_0_240312	X
349	tv_efficientnet-b1	X
350	nasnet_large	X
351	alexnet_ks_torevert	X
352	resnet18-supervised	X
353	deit_small_imagenet_1_seed-0	X
354	mobilenet_v2_0.35_160	X
355	resnet_152_v1	X
356	regnet_y_400mf	X
357	resnet_152_v2	X
358	densenet_201_pytorch	X
359	resnet50_ImageNet	X
360	vonegrcnn_62e_nobn	X
361	resnext101_32x16d_wsl	X
362	resnet_101_v2	X

Benchmark bibtex

@article{Cavanaugh2002,
            author = {Cavanaugh, James R. and Bair, Wyeth and Movshon, J. A.},
            doi = {10.1152/jn.00692.2001},
            isbn = {0022-3077 (Print) 0022-3077 (Linking)},
            issn = {0022-3077},
            journal = {Journal of Neurophysiology},
            mendeley-groups = {Benchmark effects/Done,Benchmark effects/*Surround Suppression},
            number = {5},
            pages = {2530--2546},
            pmid = {12424292},
            title = {{Nature and Interaction of Signals From the Receptive Field Center and Surround in Macaque V1 Neurons}},
            url = {http://www.physiology.org/doi/10.1152/jn.00692.2001},
            volume = {88},
            year = {2002}
            }

Ceiling

0.95.

Note that scores are relative to this ceiling.

Data: Marques2020_Cavanaugh2002

Metric: surround_diameter