Update app.py

app.py

@@ -4,14 +4,18 @@ import os
 import time
 import torch
 from scipy.io import wavfile
+import datasets
 
 # Bark imports
 from bark import generate_audio, SAMPLE_RATE
 from bark.generation import preload_models
 
 # Hugging Face Transformers
-from transformers import AutoModelForTextToSpeech, AutoProcessor, AutoTokenizer
-from transformers import SpeechT5HifiGan, SpeechT5ForTextToSpeech, SpeechT5Processor
+from transformers import (
+    SpeechT5HifiGan, 
+    SpeechT5ForTextToSpeech, 
+    SpeechT5Processor
+)
 
 class VoiceSynthesizer:
     def __init__(self):
@@ -50,7 +54,7 @@ class VoiceSynthesizer:
             vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan")
             
             # Load speaker embeddings
-            embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
+            embeddings_dataset = datasets.load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
             speaker_embeddings = torch.tensor(embeddings_dataset[0]["xvector"]).unsqueeze(0)
             
             return {