Action Retrieval on EgoExoBench
43.6ARQwen3-VL-8B + Ego4D
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-VL-8B + Ego4DModel Type=Instruct Model, Fine-tuning Data=Ego4D2026.04 | 43.6 | |
| Qwen3-VL-8B + Ego4D + DS2MModel Type=Instruct Model, Fine-tuning Data=Ego4D + DS2M2026.04 | 41.4 | |
| Qwen3-VL-emb-8BModel Type=Embedding Model2026.04 | 41.2 | |
| Qwen3-VL-emb-8B + DS2MModel Type=Embedding Model, Fine-tuning Data=DS2M2026.04 | 40.3 | |
| Qwen3-VL-8B + DS2MModel Type=Instruct Model, Fine-tuning Data=DS2M2026.04 | 38.3 | |
| Qwen3-VL-8BModel Type=Instruct Model2026.04 | 34.7 | |
| InternVL3-8BModel Type=Comparison Model2026.04 | 30.9 | |
| LLaVA-Video-7BModel Type=Comparison Model2026.04 | 29.1 |