gitroot/whisper-transcribe
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update SKILL.md thresholds and VAD integration

Browse Source
This commit is contained in:
KISEKI 2026-07-06 23:54:17 +08:00
parent f73dbeae3c
commit 8907e80451
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
SKILL.md
View File

@ -46,8 +46,8 @@ PYTHONPATH="" ~/.local/venvs/default/bin/python3 \
| **智能断句** | 修复黏连（一句字幕几秒不消失） | 检测词间停顿 >0.3s，达 `--max-line-duration` 强制切断 |
| **静音过滤** | 剔除音乐段的幽灵字幕 | ffmpeg `silencedetect` 预扫描，字幕 70% 以上在静音区则丢弃 |
| **幻觉清理** | 删除尾部复读幻觉 | 检测末尾连续重复 / 填充词 |
| **循环检测** | 修复单字 & 短句 token loop | 单字频率 + 字符串周期性检测，冷启动重转写；逆序遍历防索引漂移 [详情](references/token-repetition-loop.md) |
| **空段修复** | 补回 Whisper 漏转写的空白段 | 检测空文本段（≥2s），冷启动重转写；重试仍为空则删除 [详情](references/empty-segment-gap.md) |
| **循环检测** | 修复大段 token loop | 单字频率(≥20字) + 字符串周期(≥10次) + n-gram 去重率 + 字符密度，冷启动重转写 [详情](references/token-repetition-loop.md) |
| **空段修复** | 补回 Whisper 漏转写的大段空白 | 检测空文本段（≥15s）且非 VAD 静音区才冷启动重写；仍为空则删除 [详情](references/empty-segment-gap.md) |
## 提示词定制