Update

app.py

@@ -2,6 +2,7 @@ import gradio as gr
 from transformers import AutoModelForCausalLM, AutoProcessor, TextIteratorStreamer
 import librosa
 from threading import Thread
+import spaces
 
 def split_audio(audio_arrays, chunk_limit=480000):
     CHUNK_LIM = chunk_limit
@@ -24,7 +25,7 @@ def user(audio, text, chat_history):
     return "", chat_history
 
 
-# Placeholder for your actual LLM processing API call
+@spaces.GPU
 def process_audio(audio, text, chat_history):
     conversation = [
         {
@@ -118,7 +119,6 @@ with gr.Blocks() as demo:
         [
             ["Please transcribe the audio for me", "./examples/elon_musk.mp3"],
             ["Please transcribe the audio for me", "./examples/nvidia_conference.mp3"],
-            ["Please transcribe the audio for me", "./examples/nuggets.mp3"],
             ["Please follow the instruction in the audio", "./examples/audio_instruction.wav"],
             ["What is the primary instrument featured in the solo of this track?", "./examples/music_under.wav"],
             ["What weather condition can be heard in the audio?", "./examples/audio_understand.wav"],