Speaking Target Identification on Ego4D v1.0 (test)
66.5AccuracyRoBERTa-based Our Baseline
Evaluation Results
| Method | Links | |
|---|---|---|
| RoBERTa-based Our BaselineDensely Aligned=true2024.03 | 66.5 | |
| ELECTRA-based Our BaselineDensely Aligned=true2024.03 | 64.8 | |
| RoBERTa + MVITDensely Aligned=false2024.03 | 64.2 | |
| RoBERTa2024.03 | 63.6 | |
| RoBERTa + DINOv2Densely Aligned=false2024.03 | 62.5 | |
| BERT-based Our BaselineDensely Aligned=true2024.03 | 61.9 | |
| ELECTRA2024.03 | 60.8 | |
| ELECTRA + MVITDensely Aligned=false2024.03 | 60.8 | |
| ELECTRA + DINOv2Densely Aligned=false2024.03 | 60.2 | |
| BERT + DINOv2Densely Aligned=false2024.03 | 58 | |
| BERT + MVITDensely Aligned=false2024.03 | 57.4 | |
| BERT2024.03 | 56.8 |