List your company

Text-to-Image Retrieval on MSCOCO Caption

59.9R@1ALIGN
31.19638.64846.153.552Jun 9, 2022

Evaluation results

MethodLinks
ALIGNtuning_mode=fine-tuning (FT100%), resolution=2892022.0659.9—
Uni-Perceiver-L + Conditional MoEstuning_mode=fine-tuning (FT100%)2022.0658.3—
Uni-Perceiver-Ltuning_mode=fine-tuning (FT100%)2022.0657.9—
Uni-Perceiver-L + Conditional MoEstuning_mode=prompt tuning (PT1%)2022.0657.1—
Uni-Perceiver-Ltuning_mode=prompt tuning (PT1%)2022.0656.2—
Uni-Perceiver-L + Conditional MoEstuning_mode=w/o tuning (WT)2022.0655.3—
Uni-Perceiver-B + Conditional MoEstuning_mode=fine-tuning (FT100%)2022.0654.1—
Uni-Perceiver-Ltuning_mode=w/o tuning (WT)2022.0654.1—
Uni-Perceiver-Btuning_mode=fine-tuning (FT100%)2022.0653.9—
UNITER-Ltuning_mode=w/o tuning (WT)2022.0652.9—
Uni-Perceiver-B + Conditional MoEstuning_mode=prompt tuning (PT1%)2022.0652.6—
Uni-Perceiver-Btuning_mode=prompt tuning (PT1%)2022.0651.9—
Uni-Perceiver-B + Conditional MoEstuning_mode=w/o tuning (WT)2022.0651.6—
CoCa-Ltuning_mode=w/o tuning (WT), resolution=5762022.0651.2—
Uni-Perceiver-Btuning_mode=w/o tuning (WT)2022.0650.7—
ImageBERTtuning_mode=fine-tuning (FT100%)2022.0650.5—
UNITER-Btuning_mode=w/o tuning (WT)2022.0650.3—
Flamingo-3Btuning_mode=w/o tuning (WT), resolution=3202022.0648—
CoCa-Btuning_mode=w/o tuning (WT), resolution=5762022.0647.5—
Florencetuning_mode=w/o tuning (WT), resolution=3842022.0647.2—
ALIGNtuning_mode=w/o tuning (WT), resolution=2892022.0645.6—
ViLTtuning_mode=fine-tuning (FT100%)2022.0642.7—
ViLTtuning_mode=w/o tuning (WT)2022.0640.4—
FLAVAtuning_mode=w/o tuning (WT)2022.0638.4—
CLIPtuning_mode=w/o tuning (WT), resolution=3362022.0637.8—
ImageBERTtuning_mode=w/o tuning (WT)2022.0632.3—
CLIP-L/14Compres. ratio=0%2025.05—61.1
ReplaceMe (LS)Compres. ratio=13%2025.05—62
ReplaceMe (LS)Compres. ratio=25%2025.05—47.1
UIDLCompres. ratio=13%2025.05—60.9
UIDLCompres. ratio=25%2025.05—41.8