Text Segmentation on YTSEG (test)
52.77PrecisionMultiSeg
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MultiSegtext_encoder=fine-tuned, audio_encoder=fine-tuned2026.02 | 52.77 | 53.19 | 52.98 | 23.93 | 45.09 | |
| MiniSeg+text_encoder=fine-tuned, audio_encoder=none2026.02 | 51.87 | 46.13 | 48.83 | 25.91 | 41.17 | |
| MiniSegPre-training=WIKI-727K2024.02 | 48.3 | 43.56 | 45.81 | 27.13 | 37.89 | |
| MiniSeg + L3-Nettext_encoder=fine-tuned, audio_encoder=frozen2026.02 | 47.58 | 47.65 | 47.61 | 27.17 | 37.75 | |
| MiniSegPre-training=WIKI-727K, Future context size (c)=82024.02 | 46.92 | 41.89 | 44.26 | 27.68 | 36.81 | |
| Cross-segment BERTtext_encoder=fine-tuned, audio_encoder=none2026.02 | 46.91 | 50.02 | 48.41 | 26.47 | 40.72 | |
| MiniSegPre-training=WIKI-727K, Future context size (c)=202024.02 | 46.46 | 42.34 | 44.3 | 27.95 | 37.07 | |
| MiniSegPre-training=WIKI-727K, Future context size (c)=52024.02 | 46.24 | 42.23 | 44.15 | 27.91 | 36.62 | |
| MiniSegPre-training=WIKI-727K, Future context size (c)=32024.02 | 46.02 | 41.45 | 43.61 | 28.08 | 36.13 | |
| MiniSegPre-training=WIKI-727K, Future context size (c)=102024.02 | 45.99 | 41.31 | 43.52 | 27.97 | 36.35 | |
| MiniSegPre-training=None2024.02 | 45.44 | 41.48 | 43.37 | 28.73 | 35.74 | |
| MiniSegtext_encoder=fine-tuned, audio_encoder=none2026.02 | 45.44 | 41.48 | 43.37 | 28.73 | 35.74 | |
| MiniSegPre-training=WIKI-727K, Future context size (c)=12024.02 | 45.05 | 40.05 | 42.41 | 28.7 | 34.72 | |
| ChatGPTtext_encoder=frozen, audio_encoder=none2026.02 | 44.15 | 35.2 | 39.16 | 29.54 | 33.79 | |
| MiniSegPre-training=WIKI-727K, Future context size (c)=02024.02 | 43.69 | 37.49 | 40.35 | 29.81 | 33.11 |