Text-to-text Retrieval on COCO
31.9mAPOVI (T→I)
Evaluation results
| Method | Links | |
|---|---|---|
| OVI (T→I)Intra-modal=false, Backbone=ViT-B/16-open, Latency (ms)=11822 ± 1692026.03 | 31.9 | |
| IsoCLIPIntra-modal=true, Backbone=ViT-B/16-open, Latency (ms)=6 ± 12026.03 | 31.9 | |
| IsoCLIPIntra-modal=true, Backbone=ViT-L/14, Latency (ms)=6 ± 12026.03 | 30.1 | |
| Text-TextIntra-modal=true, Backbone=ViT-B/16-open, Latency (ms)=6 ± 12026.03 | 29.8 | |
| OVI (T→I)Intra-modal=false, Backbone=ViT-L/14, Latency (ms)=21366 ± 332026.03 | 29.4 | |
| IsoCLIPIntra-modal=true, Backbone=ViT-B/32, Latency (ms)=6 ± 12026.03 | 29.1 | |
| IsoCLIPIntra-modal=true, Backbone=ViT-B/16, Latency (ms)=6 ± 12026.03 | 28.9 | |
| OVI (T→I)Intra-modal=false, Backbone=ViT-B/32, Latency (ms)=11549 ± 1922026.03 | 28.3 | |
| OVI (T→I)Intra-modal=false, Backbone=ViT-B/16, Latency (ms)=11929 ± 1252026.03 | 28.2 | |
| Text-TextIntra-modal=true, Backbone=ViT-L/14, Latency (ms)=6 ± 12026.03 | 26.7 | |
| Text-TextIntra-modal=true, Backbone=ViT-B/32, Latency (ms)=6 ± 12026.03 | 26.2 | |
| Text-TextIntra-modal=true, Backbone=ViT-B/16, Latency (ms)=6 ± 12026.03 | 26.1 |