Referring Expression Comprehension on ReferIt (test)
81.5AccuracyPolyFormer-L
Evaluation Results
| Method | Links | |
|---|---|---|
| PolyFormer-LVisual Backbone=Swin-L, Text Encoder=BERT-base2023.02 | 81.5 | |
| PolyFormer-BVisual Backbone=Swin-B, Text Encoder=BERT-base2023.02 | 80.9 | |
| RefTrVisual Backbone=RN101, Text Encoder=BERT-base2023.02 | 76.18 | |
| SeqTRVisual Backbone=DN53, Text Encoder=Bi-GRU2023.02 | 69.66 |