Update goai_tts.py

goai_tts.py

@@ -1,11 +1,7 @@
-import torch
-import scipy
 import time
 import numpy as np
-from transformers import set_seed, pipeline
-
-device = 0 if torch.cuda.is_available() else "cpu"
-
+import scipy.io.wavfile
+from transformers import pipeline, set_seed
 
 def goai_tts(texte):
     """
@@ -18,7 +14,7 @@ def goai_tts(texte):
         
     Return
     ------
-        L'audio synthétisé.
+        Un tuple contenant le taux d'échantillonnage et les données audio sous forme de tableau numpy.
     """
     
     ### assurer la reproductibilité
@@ -28,12 +24,14 @@ def goai_tts(texte):
 
     ### charger le modèle TTS
     model_id = "anyantudre/mms-tts-mos-V1"
-    synthesiser = pipeline("text-to-speech", model_id, device=device)  # add device=0 if you want to use a GPU
+    synthesiser = pipeline("text-to-speech", model_id, device=device)
 
     ### inférence
     speech = synthesiser(texte)
-    wavfile = scipy.io.wavfile.write("finetuned_output.wav", rate=speech["sampling_rate"], data=speech["audio"][0])
+    
+    sample_rate = speech["sampling_rate"]
+    audio_data = np.array(speech["audio"][0], dtype=float)
 
     print("Temps écoulé: ", int(time.time() - start_time), " secondes")
     
-    return np.array(wavfile[1], dtype=float)
+    return sample_rate, audio_data