Text-to-Image Retrieval on DeepFashion (test) benchmark leaderboard · SOTA2 Research

Text-to-Image Retrieval on DeepFashion (test)

8.9R@1ACT-df

Evaluation Results

Method	Links
ACT-df2025.05		8.9	30.3	40.9	43.7	66.5	75.6
ACT-facad2025.05		8.7	30.2	40.7	42.8	66.1	75.4
SigLIP-ft-df2025.05		8.3	30.7	42.1	41.7	65.9	75.3
O-CLIP-ft-df2025.05		8	27.2	37.8	39.8	64	74
F-CLIP-ft-df2025.05		7.4	25.4	35	36.3	60.3	70.5
OF-CLIP-ft-df2025.05		7.2	24.5	34.4	35.8	60.7	71.2
SigLIP-ft-facad2025.05		7	25.9	36.4	35.2	56.8	68.1
O-CLIP-ft-facad2025.05		6.5	23	32.4	32.4	56.2	66.7
EVA-CLIP-ft-df2025.05		6.3	21.9	31.6	31.7	53.9	65.4
SigLIP2025.05		6.2	22.8	32.2	31.1	53.6	63.9
O-CLIP2025.05		6.1	20.4	29.1	29.8	52.7	63.7
F-CLIP2025.05		6	20.5	29.2	29.7	51.2	61.8
F-CLIP-ft-facad2025.05		6	20.6	29.5	29.5	52.6	63.1
OF-CLIP-ft-facad2025.05		5.9	20.9	29.8	29.8	54.2	64.9
OF-CLIP2025.05		5.4	18.1	26	26.8	49.1	60.6
EVA-CLIP-ft-facad2025.05		4.9	16.9	24.9	24.6	45.7	56.1
EVA-CLIP2025.05		4.5	15.1	21.8	22.5	41.2	51.7
CLIP-ft-df2025.05		2.9	9.5	14.2	14	29.9	39.3
CLIP-ft-facad2025.05		2.3	7.9	12.2	11.4	26.5	35.8
CLIP2025.05		1.9	6.1	9.1	9.2	20.1	27.8