Update app.py

app.py

@@ -1,39 +1,50 @@
-# -*- coding: utf-8 -*-
+from transformers import AutoProcessor, MusicgenForConditionalGeneration
+from IPython.display import Audio
+import scipy
+import torch
+import streamlit as st
 
-#!pip install torch
 
-#!pip install diffusers transformers accelerate torch
+def mu_gen(prompt):
+  processor = AutoProcessor.from_pretrained("facebook/musicgen-small")
+  model = MusicgenForConditionalGeneration.from_pretrained("facebook/musicgen-small")
 
-#!pip install streamlit
+  device = torch.device("cpu")
+  model.to(device)
 
-from diffusers import StableDiffusionPipeline
-import torch
-import streamlit as st
+  inputs = processor(
+      text = [str(prompt)],     # This line is correct
+      padding=True,
+      return_tensors="pt",
+  )
+  
+  inputs = {key: value.to(device) for key, value in inputs.items()}
 
-def pic_mo(prom):
-    model_id = "runwayml/stable-diffusion-v1-5"
-    pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float32)  # Change to float32
-    # No need ("cuda") if running on CPU because dont't have money 
+    # Generate audio on CPU
+  audio_values = model.generate(**inputs, max_new_tokens=256)
+  sampling_rate = model.config.audio_encoder.sampling_rate
 
-    prompt = str(prom)
+    # Create an Audio object from the generated audio
+  result = Audio(audio_values[0].numpy(), rate=sampling_rate)
 
-    image = pipe(prompt).images[0]  
-    
-    return image
+  return result
 
 
 def main():
-    st.title("Text to Image with runwayml/stable-diffusion-v1-5")
+    st.title("Text to Music Generator")
 
     # Input text prompt
-    title = st.text_input('Write a prompt (จะใช้เวลาค่อนข้างมากในการสร้างภาพเนื่องจากใช้ CPU ในการรันโมเดล)', "")
+    prompt = st.text_input("Enter a text prompt", "")
 
-    if st.button('Generate Image'):
-        # Call the pic_mo function to generate an image
-        generated_image = pic_mo(title)
+    if st.button("Generate Music"):
+        if prompt:
+            # Call the mu_gen function to generate music
+            generated_music = mu_gen(prompt)
 
-        # Display the generated image
-        st.image(generated_image, caption='Generated Image', use_column_width=True)
+            # Display the generated audio
+            st.audio(generated_music, format="audio/wav")
+        else:
+            st.warning("Please enter a text prompt.")
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     main()