Kevin676 commited on
Commit
3f89174
·
1 Parent(s): f286798

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +2 -25
app.py CHANGED
@@ -51,25 +51,14 @@ import subprocess
51
 
52
  import whisper
53
  model1 = whisper.load_model("small")
54
- os.system('pip install voicefixer --upgrade')
55
- from voicefixer import VoiceFixer
56
- voicefixer = VoiceFixer()
57
- import openai
58
- import torchaudio
59
- from speechbrain.pretrained import SpectralMaskEnhancement
60
 
61
- enhance_model = SpectralMaskEnhancement.from_hparams(
62
- source="speechbrain/metricgan-plus-voicebank",
63
- savedir="pretrained_models/metricgan-plus-voicebank",
64
- run_opts={"device":"cuda"},
65
- )
66
 
67
  mes = [
68
  {"role": "system", "content": "You are my personal assistant. Try to be helpful. Respond to me only in Chinese."}
69
  ]
70
 
71
 
72
-
73
  '''
74
  from google.colab import drive
75
  drive.mount('/content/drive')
@@ -231,19 +220,7 @@ def voice_conversion(ta, ra, da):
231
  # print("Reference Audio after decoder:")
232
  # IPython.display.display(Audio(ref_wav_voc, rate=ap.sample_rate))
233
 
234
- voicefixer.restore(input=ref_wav_voc, # input wav file path
235
- output="audio1.wav", # output wav file path
236
- cuda=True, # whether to use gpu acceleration
237
- mode = 0) # You can try out mode 0, 1, or 2 to find out the best result
238
-
239
- noisy = enhance_model.load_audio(
240
- "audio1.wav"
241
- ).unsqueeze(0)
242
-
243
- enhanced = enhance_model.enhance_batch(noisy, lengths=torch.tensor([1.]))
244
- torchaudio.save("enhanced.wav", enhanced.cpu(), 16000)
245
-
246
- return "enhanced.wav"
247
 
248
  block = gr.Blocks()
249
 
 
51
 
52
  import whisper
53
  model1 = whisper.load_model("small")
 
 
 
 
 
 
54
 
55
+ import openai
 
 
 
 
56
 
57
  mes = [
58
  {"role": "system", "content": "You are my personal assistant. Try to be helpful. Respond to me only in Chinese."}
59
  ]
60
 
61
 
 
62
  '''
63
  from google.colab import drive
64
  drive.mount('/content/drive')
 
220
  # print("Reference Audio after decoder:")
221
  # IPython.display.display(Audio(ref_wav_voc, rate=ap.sample_rate))
222
 
223
+ return (ap.sample_rate, ref_wav_voc)
 
 
 
 
 
 
 
 
 
 
 
 
224
 
225
  block = gr.Blocks()
226