Video Chaptering on VidChapters-7M (test)
45.3F1 ScoreChapter-Llama
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Chapter-LlamaBackbone=Llama 3.1-8B, Frame selection=Proposed, Finetuned=true2025.03 | 45.3 | 71.8 | 19.3 | 100.9 | |
| Gemini-1.5-ProBackbone=Gemini-1.5-Pro, Frame selection=Speech-based, Finetuned=false2025.03 | 42.2 | 70.9 | 11.4 | 63.2 | |
| Gemini-2.0-FlashBackbone=Gemini-2.0-Flash, Frame selection=Speech-based, Finetuned=false2025.03 | 40.2 | 69.3 | 11.4 | 69.7 | |
| GPT-4oBackbone=GPT-4o, Frame selection=Speech-based, Finetuned=false2025.03 | 37.6 | 68 | 8.1 | 51 | |
| GPT-4o-miniBackbone=GPT-4o-mini, Frame selection=Speech-based, Finetuned=false2025.03 | 31.2 | 63.6 | 6.8 | 37.8 | |
| Llama 3.1-8BBackbone=Llama 3.1-8B, Frame selection=Speech-based, Finetuned=false2025.03 | 29.5 | 62.5 | 6.2 | 30.7 | |
| Vid2SeqBackbone=Vid2Seq, Frame selection=Equidistant, Finetuned=true2025.03 | 26.7 | 58.6 | 11.6 | 55.8 | |
| Vid2SeqBackbone=Vid2Seq, Frame selection=Equidistant, Finetuned=false2025.03 | 3 | 29.3 | 0.3 | 0.4 |