Swiss-German-Speech-Data---Meeting-Notes.md

@@ -25,6 +25,10 @@ Yixuan Xu, Daniel Perruchoud, Michael Graber
   * Ideally for stage 2 would be daily conversations, no bias towards political parliament data
   * SRF -\> Ablation study could be done (\~ 2000 h)
   * SwissGPC? YouTube has two licences, fully permissive and fully restricted, lincences filtering could be done
+* Current timeline
+  * End of phase 1, will probably be redone
+  * Early ablations of phase 2 done
+  * Focus for now on stage 2 audio data, also TTS
 * FHNW: Transcription Evaluation for SRF broadcast transcripts
   * FHNW models perform substantially better than current SRF solution
   * ![TranscriptionRatingDistributionSRFvsFHNW.png](uploads/f4f0cc084aa9daae243cf90cb83e3485/TranscriptionRatingDistributionSRFvsFHNW.png){width="504" height="360"}