Share your thoughts, 1 month free Claude Pro on usSee more

Fine-Grained Visual Recognition on Open-world FGVR Unseen Categories

65.21Accuracy (Air)

Fine-R1-7B

Updated 3mo ago

Evaluation Results

Method
Fine-R1-7B 2026.02	65.21	60.69	82.19	70.97	40.74	82.04	66.97
Qwen2-VL-7B 2026.02	52.75	41.17	52.46	61.01	32.57	30.74	45.12
Qwen2.5-VL-3B 2026.02	52.5	48.09	51.75	59.02	34.19	28.78	45.72
Qwen2-VL-7B 2026.02	51.16	17.52	45.6	41.71	2.26	15.41	28.94
Qwen2-VL-2B 2026.02	48.88	39.32	49.33	45.16	33.87	22.43	39.83
Qwen2.5-VL-7B 2026.02	48.62	45.26	55.39	54.59	32.74	36.98	45.6
Fine-R1-3B 2026.02	46.43	58	74.6	70.08	39.54	79.11	61.29
Fine-R1-7B 2026.02	44.85	29.77	51.52	51.24	10.37	69.86	42.94
Qwen2.5-VL-7B 2026.02	44.78	28.6	45.88	49.03	10.8	27.19	34.38
Qwen2-VL-2B 2026.02	42.9	8.91	42.17	30.42	3.49	4.89	22.13
DeepPerception-7B 2026.02	42.37	21.33	47.28	48.21	5.04	40.12	34.06
Qwen2.5-VL-3B 2026.02	40.8	22.33	43.78	36.62	5.61	12.12	26.88
DeepPerception-7B 2026.02	40.03	37.1	52.27	49.05	28.38	35.57	40.4
Fine-R1-3B 2026.02	37.27	30.68	45.73	49.79	16.5	61.02	40.17
InternVL2.5-8B 2026.02	20.34	24.38	27.55	24.73	13.56	26.23	22.8
InternVL2.5-4B 2026.02	12.64	23.91	29.11	30.8	13.1	26.32	22.65
InternVL2.5-8B 2026.02	8.64	9	8.42	11.53	1.79	13.66	8.84
LLaVA-Onevision-7B 2026.02	7.47	29.81	19.56	25.44	16.53	19.35	19.69
Idefics3-LLaMA3-8B 2026.02	7.09	24.93	22.08	27.67	21.99	24.84	21.43
InternVL2.5-4B 2026.02	7.06	7.1	10.96	11.2	1.84	10.11	8.05
Idefics2-8B 2026.02	6.76	6.01	13.37	12.14	0.38	11.45	8.35
InternVL2.5-2B 2026.02	5.22	20.12	24.73	24.84	12.59	24.55	18.68
Idefics2-8B 2026.02	3.69	15.35	20.81	10.19	5.84	2.47	9.73
LLaVA-Onevision-7B 2026.02	3.46	5.54	15.36	13.01	0.14	2.68	6.7
Idefics3-LLaMA3-8B 2026.02	3.16	3.5	2.97	9.59	2.12	7.57	4.82
LLaVA-v1.6-mistral-7B 2026.02	2.51	17.4	23.24	18.16	8.2	10.62	13.36
InternVL2.5-2B 2026.02	2.33	4.2	7.24	9.41	1.27	13.06	6.25
LLaVA-v1.6-mistral-7B 2026.02	2.03	3.25	8.39	9.59	0.09	7.97	5.22