AlexK-PL commited on
Commit
c8797aa
verified
1 Parent(s): b0910d2

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +19 -10
app.py CHANGED
@@ -57,17 +57,17 @@ commonvoice_speakers = [s for s in speakers_list if len(s) > 20] #
57
 
58
  DEFAULT_SPEAKER_ID = os.environ.get("DEFAULT_SPEAKER_ID", default="pau")
59
  DEFAULT_CHECKPOINT = os.environ.get("DEFAULT_CHECKPOINT", default=model_files[0])
60
- model_file = model_files[0] # change this!!
61
 
62
- model_path = os.path.join(os.getcwd(), model_file)
63
- config_path = "config.json"
64
 
65
  vocoder_path = None
66
  vocoder_config_path = None
67
 
68
- synthesizer = Synthesizer(
69
- model_path, config_path, speakers_path, None, vocoder_path, vocoder_config_path,
70
- )
71
 
72
 
73
  def get_phonetic_transcription(text: str):
@@ -85,7 +85,16 @@ def get_phonetic_transcription(text: str):
85
  return None
86
 
87
 
88
- def tts_inference(text: str, speaker_idx: str = None):
 
 
 
 
 
 
 
 
 
89
  # synthesize
90
  if synthesizer is None:
91
  raise NameError("model not found")
@@ -113,11 +122,11 @@ description = """
113
  """
114
 
115
 
116
- def submit_input(input_, speaker_id):
117
  output_audio = None
118
  output_phonetic = None
119
  if input_ is not None and len(input_) < MAX_INPUT_TEXT_LEN:
120
- output_audio = tts_inference(input_, speaker_id)
121
  output_phonetic = get_phonetic_transcription(input_)
122
  else:
123
  gr.Warning(f"Your text exceeds the {MAX_INPUT_TEXT_LEN}-character limit.")
@@ -196,7 +205,7 @@ with gr.Blocks(**AinaGradioTheme().get_kwargs()) as app:
196
  input_.change(fn=change_interactive, inputs=[input_], outputs=button)
197
 
198
  # clear_btn.click(fn=clean, inputs=[], outputs=[input_, output_audio, output_phonetic], queue=False)
199
- submit_btn.click(fn=submit_input, inputs=[input_, speaker_id], outputs=[output_audio, output_phonetic])
200
 
201
  app.queue(concurrency_count=1, api_open=False)
202
  app.launch(show_api=False, server_name="0.0.0.0", server_port=7860)
 
57
 
58
  DEFAULT_SPEAKER_ID = os.environ.get("DEFAULT_SPEAKER_ID", default="pau")
59
  DEFAULT_CHECKPOINT = os.environ.get("DEFAULT_CHECKPOINT", default=model_files[0])
60
+ # model_file = model_files[0] # change this!!
61
 
62
+ # model_path = os.path.join(os.getcwd(), model_file)
63
+ # config_path = os.path.join(os.getcwd(), "config.json")
64
 
65
  vocoder_path = None
66
  vocoder_config_path = None
67
 
68
+ # synthesizer = Synthesizer(
69
+ # model_path, config_path, speakers_path, None, vocoder_path, vocoder_config_path,
70
+ # )
71
 
72
 
73
  def get_phonetic_transcription(text: str):
 
85
  return None
86
 
87
 
88
+ def tts_inference(text: str, speaker_idx: str = None, model_file: str=None):
89
+
90
+ model_path = os.path.join(os.getcwd(), model_file)
91
+ speakers_file_path = "speakers.pth"
92
+ config_path = "config.json"
93
+ vocoder_path = None
94
+ vocoder_config_path = None
95
+
96
+ synthesizer = Synthesizer(model_path, config_path, speakers_path, None,
97
+ vocoder_path, vocoder_config_path)
98
  # synthesize
99
  if synthesizer is None:
100
  raise NameError("model not found")
 
122
  """
123
 
124
 
125
+ def submit_input(input_, speaker_id, model_chkpt):
126
  output_audio = None
127
  output_phonetic = None
128
  if input_ is not None and len(input_) < MAX_INPUT_TEXT_LEN:
129
+ output_audio = tts_inference(input_, speaker_id, model_chkpt)
130
  output_phonetic = get_phonetic_transcription(input_)
131
  else:
132
  gr.Warning(f"Your text exceeds the {MAX_INPUT_TEXT_LEN}-character limit.")
 
205
  input_.change(fn=change_interactive, inputs=[input_], outputs=button)
206
 
207
  # clear_btn.click(fn=clean, inputs=[], outputs=[input_, output_audio, output_phonetic], queue=False)
208
+ submit_btn.click(fn=submit_input, inputs=[input_, speaker_id, model_chkpt], outputs=[output_audio, output_phonetic])
209
 
210
  app.queue(concurrency_count=1, api_open=False)
211
  app.launch(show_api=False, server_name="0.0.0.0", server_port=7860)