Update app.py

app.py

@@ -10,6 +10,11 @@ import uuid
 from stable_audio_tools import get_pretrained_model
 from stable_audio_tools.inference.generation import generate_diffusion_cond
 
+# Load the model outside of the GPU-decorated function
+def load_model():
+    model, model_config = get_pretrained_model("stabilityai/stable-audio-open-1.0")
+    return model, model_config
+
 # Function to set up, generate, and process the audio
 @spaces.GPU(duration=120)  # Allocate GPU only when this function is called
 def generate_audio(prompt, seconds_total=30, steps=100, cfg_scale=7):
@@ -17,9 +22,9 @@ def generate_audio(prompt, seconds_total=30, steps=100, cfg_scale=7):
 
     # Fetch the Hugging Face token from the environment variable
     hf_token = os.getenv('HF_TOKEN')
-    
-    # Download and set up the model
-    model, model_config = get_pretrained_model("stabilityai/stable-audio-open-1.0")
+
+    # Use pre-loaded model and configuration
+    model, model_config = load_model()
     sample_rate = model_config["sample_rate"]
     sample_size = model_config["sample_size"]
 
@@ -73,5 +78,8 @@ interface = gr.Interface(
     description="Generate variable-length stereo audio at 44.1kHz from text prompts using Stable Audio Open 1.0."
 )
 
+# Pre-load the model to avoid multiprocessing issues
+model, model_config = load_model()
+
 # Launch the Interface
 interface.launch()