tipi/kyutai
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Allow for using local voices in the pytorch examples. (#100)

Browse Source
This commit is contained in:
Laurent Mazare 2025-07-31 12:48:05 +02:00 committed by GitHub
parent ab8e8c59b7
commit 7dc926d50c
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 8 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
scripts/tts_pytorch.py
View File

@ -68,6 +68,9 @@ def main():
# If you want to make a dialog, you can pass more than one turn [text_speaker_1, text_speaker_2, text_2_speaker_1, ...] # If you want to make a dialog, you can pass more than one turn [text_speaker_1, text_speaker_2, text_2_speaker_1, ...]
entries = tts_model.prepare_script([text], padding_between=1) entries = tts_model.prepare_script([text], padding_between=1)
if args.voice.endswith(".safetensors"):
voice_path = args.voice
else:
voice_path = tts_model.get_voice_path(args.voice) voice_path = tts_model.get_voice_path(args.voice)
# CFG coef goes here because the model was trained with CFG distillation, # CFG coef goes here because the model was trained with CFG distillation,
# so it's not _actually_ doing CFG at inference time. # so it's not _actually_ doing CFG at inference time.

3
scripts/tts_pytorch_streaming.py
View File

@ -183,6 +183,9 @@ def main():
checkpoint_info, n_q=32, temp=0.6, device=args.device checkpoint_info, n_q=32, temp=0.6, device=args.device
) )
if args.voice.endswith(".safetensors"):
voice_path = args.voice
else:
voice_path = tts_model.get_voice_path(args.voice) voice_path = tts_model.get_voice_path(args.voice)
# CFG coef goes here because the model was trained with CFG distillation, # CFG coef goes here because the model was trained with CFG distillation,
# so it's not _actually_ doing CFG at inference time. # so it's not _actually_ doing CFG at inference time.