Ethanfel a0286d5cf9 feat: waveform overlay, signal safety, training cancel, dynamic batch size, duplicate detection ...

- WaveformWorker extracts low-res audio envelope via ffmpeg, drawn as
  green polygon on timeline track
- _safe_disconnect() replaces bare TypeError catches for signal cleanup
- Train button toggles to Cancel during training, calls worker.cancel()
- Dynamic GPU batch sizing: 64 for ≥16GB VRAM, 32 for ≥8GB, 16 default
- Overlap warning before exporting clips that intersect existing markers

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2026-04-19 12:53:48 +02:00

..

__init__.py

feat: create core/paths module with shared path helpers

2026-04-16 13:34:17 +02:00

annotations.py

feat: create core/annotations module

2026-04-16 13:38:47 +02:00

audio_scan.py

feat: waveform overlay, signal safety, training cancel, dynamic batch size, duplicate detection

2026-04-19 12:53:48 +02:00

beats_backbone.py

feat: integrate training UI, BEATs model, and clean up legacy code

2026-04-18 11:52:27 +02:00

beats_model.py

feat: integrate training UI, BEATs model, and clean up legacy code

2026-04-18 11:52:27 +02:00

beats_modules.py

feat: integrate training UI, BEATs model, and clean up legacy code

2026-04-18 11:52:27 +02:00

db.py

fix: DB schema missing scan_export column, add threshold filter and N hotkey

2026-04-19 12:45:14 +02:00

ffmpeg.py

feat: disable/resize scan regions, undo, training fixes, cross-platform cleanup

2026-04-18 20:34:56 +02:00

paths.py

feat: create core/paths module with shared path helpers

2026-04-16 13:34:17 +02:00

tracking.py

feat: create core/tracking module with YOLO subject tracking

2026-04-16 13:39:52 +02:00