jason-on-salt-a40 commited on
Commit
0d18e61
1 Parent(s): cf33c41

better hyperparams

Browse files
Files changed (1) hide show
  1. app.py +3 -3
app.py CHANGED
@@ -460,7 +460,7 @@ def get_app():
460
 
461
  with gr.Group() as tts_mode_controls:
462
  prompt_to_word = gr.Dropdown(label="Last word in prompt", choices=demo_words, value=demo_words[12], interactive=True)
463
- prompt_end_time = gr.Slider(label="Prompt end time", minimum=0, maximum=7.479, step=0.001, value=3.700)
464
 
465
  with gr.Group(visible=False) as edit_mode_controls:
466
  with gr.Row():
@@ -485,9 +485,9 @@ def get_app():
485
 
486
  with gr.Row():
487
  with gr.Accordion("Generation Parameters - change these if you are unhappy with the generation", open=False):
488
- stop_repetition = gr.Radio(label="stop_repetition", choices=[-1, 1, 2, 3, 4], value=2,
489
  info="if there are long silence in the generated audio, reduce the stop_repetition to 2 or 1. -1 = disabled")
490
- sample_batch_size = gr.Number(label="speech rate", value=4, precision=0,
491
  info="The higher the number, the faster the output will be. "
492
  "Under the hood, the model will generate this many samples and choose the shortest one. "
493
  "For giga330M_TTSEnhanced, 1 or 2 should be fine since the model is trained to do TTS.")
 
460
 
461
  with gr.Group() as tts_mode_controls:
462
  prompt_to_word = gr.Dropdown(label="Last word in prompt", choices=demo_words, value=demo_words[12], interactive=True)
463
+ prompt_end_time = gr.Slider(label="Prompt end time", minimum=0, maximum=7.479, step=0.001, value=3.689)
464
 
465
  with gr.Group(visible=False) as edit_mode_controls:
466
  with gr.Row():
 
485
 
486
  with gr.Row():
487
  with gr.Accordion("Generation Parameters - change these if you are unhappy with the generation", open=False):
488
+ stop_repetition = gr.Radio(label="stop_repetition", choices=[-1, 1, 2, 3, 4], value=3,
489
  info="if there are long silence in the generated audio, reduce the stop_repetition to 2 or 1. -1 = disabled")
490
+ sample_batch_size = gr.Number(label="speech rate", value=2, precision=0,
491
  info="The higher the number, the faster the output will be. "
492
  "Under the hood, the model will generate this many samples and choose the shortest one. "
493
  "For giga330M_TTSEnhanced, 1 or 2 should be fine since the model is trained to do TTS.")