Image Captioning on COCO Caption (VQA/Reasoning Evaluation Set) benchmark leaderboard · SOTA2 Research

Image Captioning on COCO Caption (VQA/Reasoning Evaluation Set)

57.99OKVQAZero-shot

Evaluation Results

Method	Links
Zero-shot2025.03		57.99	66.2	61.93	58.23	61.09	40.4	50.74
Model Tailor2025.03		51.83	59.75	58.73	51.43	55.44	118.84	87.14
Model Tailor2025.03		50.99	60.6	58.54	49.65	54.95	117.64	86.29
LoRASculpt2025.03		50.86	59.1	57.64	51.1	54.68	121.26	87.97
LoRASculpt2025.03		50.15	59.85	57.91	51.44	54.84	120.39	87.61
LoRASculpt2025.03		49.99	58.65	57.63	50.73	54.25	120.35	87.3
Orth-Reg2025.03		48.11	59.85	56.83	46.69	52.87	111.87	82.37
LoRA2025.03		47.34	60.55	56.91	45.61	52.6	112.22	82.41
Model Tailor2025.03		46.7	59.2	56.75	44.42	51.77	116.68	84.22
DARE2025.03		46.64	59.45	56.01	44.46	51.64	112.21	81.93
L2-Reg2025.03		46.05	61.5	56.6	43.54	51.92	111.83	81.88
Orth-Reg2025.03		45.78	60.7	55.69	43.58	51.44	108.16	79.8
DoRA2025.03		45.38	59.75	55.4	41.38	50.48	112.6	81.54
LoRA2025.03		44.92	59.9	54.88	39.75	49.86	110.27	80.07
DoRA2025.03		44.1	60.05	54.69	40.17	49.75	109.25	79.5
Orth-Reg2025.03		43.77	60.7	53.74	38.61	49.21	110.44	79.82
L2-Reg2025.03		43.6	60	54.09	36.53	48.56	109.39	78.97
DARE2025.03		42.98	58.65	53.92	38.61	48.54	108.57	78.56
L2-Reg2025.03		40.95	59.05	52.83	36	47.21	106.41	76.81
DoRA2025.03		34.63	57.85	44.37	29.06	41.48	106.49	73.98
LoRA2025.03		31.51	54.8	42.35	28.13	39.2	107.9	73.55
DARE2025.03		29.67	50.7	37.54	23.34	35.31	105.55	70.43