init

app.py

@@ -6,6 +6,8 @@ import soundfile as sf
 import torch
 from transformers import AutoModelForCausalLM, AutoProcessor, GenerationConfig
 
+import spaces
+
 
 MODEL_ID = "microsoft/Phi-4-multimodal-instruct"
 
@@ -20,7 +22,7 @@ model = AutoModelForCausalLM.from_pretrained(
 
 generation_config = GenerationConfig.from_pretrained(MODEL_ID)
 
-
+@spaces.GPU
 def run_phi4(audio_path: str, instruction: str, max_tokens: int = 512) -> str:
     if not audio_path:
         return "Please upload an audio file."
@@ -68,12 +70,4 @@ with gr.Blocks(title="Phi-4 Multimodal Audio Demo") as demo:
 
 
 if __name__ == "__main__":
-    demo.queue().launch(share=False, ssr_mode=False)
-
-import gradio as gr
-
-def greet(name):
-    return "Hello " + name + "!!"
-
-demo = gr.Interface(fn=greet, inputs="text", outputs="text")
-demo.launch()
+    demo.queue().launch(share=False, ssr_mode=False)

requirements.txt

@@ -4,4 +4,5 @@ transformers>=4.43.0
 accelerate>=0.30.0
 soundfile>=0.12.1
 sentencepiece>=0.1.99
+spaces