Update app.py

app.py

@@ -3,6 +3,7 @@ import torch
 from datasets import load_dataset
 from transformers import SpeechT5Processor, SpeechT5ForTextToSpeech, SpeechT5HifiGan
 import soundfile as sf
+import numpy as np
 
 # Load the fine-tuned model, processor, and vocoder
 model_name = "microsoft/speecht5_tts"
@@ -10,16 +11,15 @@ processor = SpeechT5Processor.from_pretrained(model_name)
 model = SpeechT5ForTextToSpeech.from_pretrained("emirhanbilgic/speecht5_finetuned_emirhan_tr")
 vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan")
 
-# Load speaker embeddings (using the same as before, as it's not clear if a specific embedding is needed for the Turkish model)
+# Load speaker embeddings
 embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
 speaker_embeddings = torch.tensor(embeddings_dataset[7306]["xvector"]).unsqueeze(0)
 
-
 def text_to_speech(text):
     inputs = processor(text=text, return_tensors="pt")
     speech = model.generate_speech(inputs["input_ids"], speaker_embeddings, vocoder=vocoder)
-    sf.write("output.wav", speech.numpy(), samplerate=16000)
-    return "output.wav"
+    speech_numpy = speech.numpy()
+    return (16000, speech_numpy)  # Return sample rate and numpy array
 
 # Create Gradio interface
 iface = gr.Interface(