Update app.py

app.py

@@ -1,46 +1,23 @@
-from transformers import pipeline
-from datasets import load_dataset
-import soundfile as sf
-import torch
 import gradio as gr
-import os
+from bark import SAMPLE_RATE, generate_audio, preload_models
+from IPython.display import Audio
 
-# Text-to-speech pipeline
-synthesiser = pipeline("text-to-speech", "microsoft/speecht5_tts")
-embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
+# Preload models if needed
+preload_models()
 
-def synthesize_speech(text, speaker_id):
-    try:
-        if not text.strip():
-            return "Please enter valid text."
+def create_audio(text):
+    audio_array = generate_audio(text)
+    return audio_array, SAMPLE_RATE
 
-        speaker_embedding = torch.tensor(embeddings_dataset[speaker_id]["xvector"]).unsqueeze(0)
-        speech = synthesiser(text, forward_params={"speaker_embeddings": speaker_embedding})
-
-        # Saving as WAV file
-        wav_file_path = "speech.wav"
-        sf.write(wav_file_path, speech["audio"], samplerate=speech["sampling_rate"])
-
-        # Converting to MP3
-        mp3_file_path = "speech.mp3"
-        os.system(f"ffmpeg -i {wav_file_path} -acodec libmp3lame {mp3_file_path}")
-
-        return mp3_file_path
-    except Exception as e:
-        return f"An error occurred: {str(e)}"
-
-
-# Improved Interface
 iface = gr.Interface(
-    fn=synthesize_speech,
-    inputs=[
-        gr.Textbox(label="Enter Text", placeholder="Type here to practice English pronunciation..."),
-        gr.Slider(minimum=0, maximum=len(embeddings_dataset)-1, label="Select Speaker Voice")
+    fn=create_audio,
+    inputs=gr.inputs.Textbox(lines=5, placeholder="Digite seu texto aqui..."),
+    outputs=[
+        gr.outputs.Audio(type="numpy", label="Audio Gerado"),
+        gr.outputs.Textbox(label="Taxa de Amostragem")
     ],
-    outputs=gr.Audio(label="Pronunciation Audio", type="filepath"),
-    title="English Pronunciation Helper",
-    description="This tool helps you practice English pronunciation. Type any text in English, and hear how it's pronounced. You can also download the audio in MP3 format."
+    title="Gerador de Áudio com IA",
+    description="Digite um texto para gerar áudio usando IA."
 )
 
-
 iface.launch()