Hugo Rodrigues commited on
Commit
b2b9472
·
1 Parent(s): 6db451f

remove dependency torchaudio

Browse files
Files changed (1) hide show
  1. main.py +7 -7
main.py CHANGED
@@ -1,7 +1,6 @@
1
 
2
  import time
3
  from scipy.io.wavfile import write
4
- import torchaudio
5
  import numpy as np
6
 
7
 
@@ -88,21 +87,19 @@ async def audio(inputs, src_lang="eng", tgt_lang="por", speaker_id=5):
88
  audio_array_from_text = model.generate(
89
  **text_inputs, tgt_lang=tgt_lang, speaker_id=int(speaker_id))[0].cpu().numpy().squeeze()
90
 
91
- print("Time took to process the request and return response is {} sec".format(
92
- time.time() - start_time))
93
-
94
- print(f"sampling_rate {model.config.sampling_rate}")
95
-
96
  write(f"/tmp/output{start_time}.wav", model.config.sampling_rate,
97
  audio_array_from_text)
98
 
 
 
 
99
  return FileResponse(f"/tmp/output{start_time}.wav", media_type="audio/mpeg")
100
 
101
 
102
  @app.post("/transcribe-audio")
103
  async def transcribe_audio(soundFile: UploadFile, tgt_lang='eng'):
104
  start_time = time.time()
105
- # process input
106
  inputFile = soundFile.file.read()
107
 
108
  audio_data = np.frombuffer(inputFile, dtype=np.int16)
@@ -116,4 +113,7 @@ async def transcribe_audio(soundFile: UploadFile, tgt_lang='eng'):
116
  write(f"/tmp/output{start_time}.wav", model.config.sampling_rate,
117
  audio_array_from_audio)
118
 
 
 
 
119
  return FileResponse(f"/tmp/output{start_time}.wav", media_type="audio/wav")
 
1
 
2
  import time
3
  from scipy.io.wavfile import write
 
4
  import numpy as np
5
 
6
 
 
87
  audio_array_from_text = model.generate(
88
  **text_inputs, tgt_lang=tgt_lang, speaker_id=int(speaker_id))[0].cpu().numpy().squeeze()
89
 
 
 
 
 
 
90
  write(f"/tmp/output{start_time}.wav", model.config.sampling_rate,
91
  audio_array_from_text)
92
 
93
+ print("Time took to process the request and return response is {} sec".format(
94
+ time.time() - start_time))
95
+
96
  return FileResponse(f"/tmp/output{start_time}.wav", media_type="audio/mpeg")
97
 
98
 
99
  @app.post("/transcribe-audio")
100
  async def transcribe_audio(soundFile: UploadFile, tgt_lang='eng'):
101
  start_time = time.time()
102
+
103
  inputFile = soundFile.file.read()
104
 
105
  audio_data = np.frombuffer(inputFile, dtype=np.int16)
 
113
  write(f"/tmp/output{start_time}.wav", model.config.sampling_rate,
114
  audio_array_from_audio)
115
 
116
+ print("Time took to process the request and return response is {} sec".format(
117
+ time.time() - start_time))
118
+
119
  return FileResponse(f"/tmp/output{start_time}.wav", media_type="audio/wav")