Image Classification on 12 Downstream Datasets (Fine-grained, Natural, Specialized) (test) benchmark leaderboard · SOTA2 Research

Image Classification on 12 Downstream Datasets (Fine-grained, Natural, Specialized) (test)

72.07Accuracy (Fine-grained)IOTA

Evaluation Results

Method	Links
IOTA2026.01		72.07	72.24	85.87	76.73
ProText2026.01		67.39	68.06	83.44	72.96
CoOp2026.01		65.95	64.68	82.86	71.16
Clip-Adapter2026.01		65.85	63.77	83.85	71.16
TCP2026.01		65.14	67.73	82.91	71.93
IOTA2026.01		64.75	67.17	82.82	71.58
LP++2026.01		64.65	66.63	82.48	71.25
Maple2026.01		64.49	68.37	83.32	72.06
KgCoOp2026.01		64.23	65.92	82.11	70.75
ProGrad2026.01		63.91	64.48	79.52	69.3
ProText2026.01		62.66	64.15	79.46	68.76
LP++2026.01		62.27	62.85	79.7	68.27
CoOp2026.01		61.37	62.17	78.19	67.24
ProGrad2026.01		60.97	62.59	74.02	65.86
Clip-Adapter2026.01		60.91	62.72	79.05	67.56
TCP2026.01		60.73	63.81	78.97	67.84
KgCoOp2026.01		60.15	64.02	79.63	67.93
Maple2026.01		60.01	65.47	77.3	67.59
Co-CoOp2026.01		59.41	63.2	78.65	67.09
Co-CoOp2026.01		55.91	62.08	74.23	64.08
CLIP2026.01		49.96	63.86	51.9	55.24