Update app.py

app.py

@@ -6,52 +6,51 @@ import logging
 import tempfile
 import numpy as np
 
-# Configura o logging
+# Configure logging
 logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
 
 @spaces.GPU(duration=120)
 def generate_music(description, melody_audio):
-    logging.info("Iniciando a geração de música.")
+    logging.info("Starting music generation.")
 
-    # Carrega o modelo pré-treinado
-    logging.info("Carregando o modelo pré-treinado.")
+    # Load the pre-trained model
+    logging.info("Loading pre-trained model.")
     model = MusicGen.get_pretrained('nateraw/musicgen-songstarter-v0.2')
     model.set_generation_params(duration=8)
 
     if description:
         description = [description]
         if melody_audio:
-            logging.info(f"Carregando a melodia de áudio de: {melody_audio}")
+            logging.info(f"Loading audio melody from: {melody_audio}")
             melody, sr = torchaudio.load(melody_audio)
-            logging.info("Gerando música com descrição e melodia.")
+            logging.info("Generating music with description and melody.")
             wav = model.generate_with_chroma(description, melody[None], sr)
         else:
-            logging.info("Gerando música apenas com descrição.")
+            logging.info("Generating music with description only.")
             wav = model.generate(description)
     else:
-        logging.info("Gerando música de forma incondicional.")
+        logging.info("Generating music unconditionally.")
         wav = model.generate_unconditional(1)
 
-    # Verifica a forma do tensor de áudio gerado
-    logging.info(f"A forma do tensor de áudio gerado: {wav[0].shape}")
-    logging.info("Música gerada com sucesso.")
-
-    # Retorna o áudio gerado como uma tupla (sample rate, numpy array)
-    return model.sample_rate, wav[0].cpu().numpy()
-
-# Define a interface Gradio
-description = gr.Textbox(label="Description", placeholder="acoustic, guitar, melody, trap, d minor, 90 bpm")
-melody_audio = gr.Audio(label="Melody Audio (optional)", type="filepath")
-output_audio = gr.Audio(label="Generated Music", type="numpy")
-
-gr.Interface(
-    fn=generate_music,
-    inputs=[description, melody_audio],
-    outputs=output_audio,
-    title="MusicGen Demo",
-    description="Generate music using the MusicGen model.",
-    examples=[
-        ["trap, synthesizer, songstarters, dark, G# minor, 140 bpm", "./assets/kalhonaho.mp3"],
-        ["upbeat, electronic, synth, dance, 120 bpm", None]
-    ]
-).launch()
+    # Check the shape of the generated audio tensor
+    logging.info(f"The shape of the generated audio tensor: {wav[0].shape}")
+
+    # Convert the generated audio to the format expected by gr.Audio
+    audio_data = (model.sample_rate, (wav[0].cpu().numpy() * 32767).astype(np.int16))
+    logging.info("Music generated successfully.")
+
+    return audio_data
+
+# Create the Gradio interface
+with gr.Blocks() as demo:
+    with gr.Row():
+        with gr.Column():
+            description = gr.Textbox(label="Music Description")
+            melody_audio = gr.Audio(label="Melody Audio")
+        with gr.Column():
+            output_audio = gr.Audio(label="Generated Music", interactive=False)
+
+    generate_btn = gr.Button("Generate Music")
+    generate_btn.click(generate_music, inputs=[description, melody_audio], outputs=[output_audio])
+
+demo.launch()