Create app.py

app.py

@@ -0,0 +1,43 @@
+import gradio as gr
+import torch
+from diffusers import AudioLDMPipeline
+
+print("Loading AudioLDM Model (This takes a few minutes on boot)...")
+# Using the small model to fit in the free tier RAM
+repo_id = "cvssp/audioldm-s-full-v2"
+pipe = AudioLDMPipeline.from_pretrained(repo_id, torch_dtype=torch.float32)
+
+def generate_audio(text_prompt, duration, guidance):
+    print(f"Generating sound for: {text_prompt}")
+    # Number of inference steps reduced slightly for faster CPU generation
+    audio = pipe(
+        text_prompt, 
+        num_inference_steps=15, 
+        audio_length_in_s=duration, 
+        guidance_scale=guidance
+    ).audios[0]
+    
+    # Diffusers outputs audio at 16000Hz sampling rate
+    return (16000, audio)
+
+with gr.Blocks(theme=gr.themes.Monochrome()) as app:
+    gr.Markdown("# 🎬 The Foley Artist AI")
+    gr.Markdown("Generate high-fidelity sound effects from text descriptions.")
+    
+    with gr.Row():
+        with gr.Column():
+            prompt = gr.Textbox(label="Describe the Sound (e.g., 'Footsteps on gravel')", lines=2)
+            duration = gr.Slider(minimum=2.5, maximum=5.0, value=2.5, step=2.5, label="Duration (Seconds)")
+            guidance = gr.Slider(minimum=1.0, maximum=5.0, value=2.5, step=0.5, label="Guidance Scale (Higher = closer to text)")
+            generate_btn = gr.Button("Generate Audio", variant="primary")
+            
+        with gr.Column():
+            audio_output = gr.Audio(label="Generated Sound Effect")
+            
+    generate_btn.click(
+        generate_audio, 
+        inputs=[prompt, duration, guidance], 
+        outputs=[audio_output]
+    )
+
+app.launch()