shaikrabbani-dev commited on
Commit
6f9caf2
·
verified ·
1 Parent(s): 36a0cd0

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +17 -8
app.py CHANGED
@@ -78,25 +78,34 @@ def clone_and_speak(text):
78
  unique_filename = f"voice_output_{uuid.uuid4().hex[:8]}.wav"
79
  print(f"🔊 Rendering voice cloning track for text: '{text}'")
80
 
81
- # FIXED: Enforced a speed factor and reduced timesteps to guarantee execution under CPU limitations
82
- audio_data, sample_rate, _ = f5tts.infer(
83
  ref_file=voice_file,
84
  ref_text=auto_reference_text,
85
- gen_text=text,
86
- n_timesteps=10, # Drastically speeds up synthesis step count
87
- speed=1.1 # Compresses wave calculations slightly for speed
88
  )
89
 
90
- # Ensure it is explicitly a clean numpy array for soundfile
 
 
 
 
 
 
 
 
 
 
 
91
  if hasattr(audio_data, "cpu"):
92
  audio_data = audio_data.cpu().numpy()
93
  elif hasattr(audio_data, "numpy"):
94
  audio_data = audio_data.numpy()
95
 
96
- # Ensure array dimensions are flattened correctly
97
  audio_data = np.squeeze(audio_data)
98
 
99
- # Write to absolute target path
100
  sf.write(unique_filename, audio_data, sample_rate, format='WAV')
101
 
102
  if os.path.exists(unique_filename) and os.path.getsize(unique_filename) > 0:
 
78
  unique_filename = f"voice_output_{uuid.uuid4().hex[:8]}.wav"
79
  print(f"🔊 Rendering voice cloning track for text: '{text}'")
80
 
81
+ # Invoke inference engine
82
+ result = f5tts.infer(
83
  ref_file=voice_file,
84
  ref_text=auto_reference_text,
85
+ gen_text=text
 
 
86
  )
87
 
88
+ # CRITICAL FIX: Dynamically detect and unpack tuple structures regardless of package version variations
89
+ if isinstance(result, tuple):
90
+ if len(result) == 2:
91
+ audio_data, sample_rate = result
92
+ elif len(result) >= 3:
93
+ audio_data, sample_rate = result[0], result[1]
94
+ else:
95
+ audio_data, sample_rate = result[0], 24000
96
+ else:
97
+ audio_data, sample_rate = result, 24000
98
+
99
+ # Convert Torch tensors safely to standard numpy array structures
100
  if hasattr(audio_data, "cpu"):
101
  audio_data = audio_data.cpu().numpy()
102
  elif hasattr(audio_data, "numpy"):
103
  audio_data = audio_data.numpy()
104
 
105
+ # Flatten structural multi-dimensional arrays safely for soundfile
106
  audio_data = np.squeeze(audio_data)
107
 
108
+ # Write clean output wave file natively
109
  sf.write(unique_filename, audio_data, sample_rate, format='WAV')
110
 
111
  if os.path.exists(unique_filename) and os.path.getsize(unique_filename) > 0: