Trying ray

app.py

@@ -1,7 +1,7 @@
 import spaces
 import gradio as gr
 import torch
-
+import ray
 
 
 # Use a pipeline as a high-level helper
@@ -13,6 +13,7 @@ print("RUNNING PIPE")
 pipe = pipeline("text-generation", model="NousResearch/Hermes-3-Llama-3.1-8B", max_new_tokens=200, device=0)
 print("PIPE DONE") 
 
+@ray.remote(num_gpus=1, max_calls=1)
 @spaces.GPU(duration=120)
 def llama3_1_8B(question):