local inference page with fixes to gpu with zerogpu

app.py

@@ -11,6 +11,7 @@ from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer, TextIter
 from threading import Thread
 from typing import Iterator
 import os
+import spaces
 
 # Set up logging
 logging.basicConfig(level=logging.INFO)
@@ -110,6 +111,7 @@ def analyze_report(user_findings, ground_truth_findings, ground_truth_impression
         logger.error(f"Error in report analysis: {str(e)}")
         return f"Error analyzing report: {str(e)}"
 
+@spaces.GPU
 def transcribe(inputs, task="transcribe"):
     """Transcribe audio using Whisper"""
     if inputs is None:
@@ -125,6 +127,7 @@ def transcribe(inputs, task="transcribe"):
         logger.error(f"Error in transcription: {str(e)}")
         raise gr.Error(f"Transcription failed: {str(e)}")
 
+@spaces.GPU
 def analyze_with_llama(
     transcribed_text: str,
     ground_truth_findings: str,