add gput

app.py

@@ -46,6 +46,7 @@ tokenizer, model, data_loader = load_model_and_dataloader(
     scene_to_obj_mapping=scene_to_obj_mapping,
 )
 
+@spaces.GPU
 def get_chatbot_response(user_chat_input, scene_id):
     # Get the response from the model
     prompt, response = get_model_response(

model.py

@@ -1,4 +1,5 @@
 # model.py
+import spaces
 import json
 import torch
 from torch.utils.data import DataLoader
@@ -11,7 +12,7 @@ from llava.mm_utils import get_model_name_from_path
 from llava.model.builder import load_pretrained_model
 from llava.utils import disable_torch_init
 
-
+@spaces.GPU
 def load_model_and_dataloader(model_path, model_base, scene_to_obj_mapping, obj_context_feature_type="text", load_8bit=False, load_4bit=False, load_bf16=False):
     disable_torch_init()
 
@@ -41,6 +42,7 @@ def load_model_and_dataloader(model_path, model_base, scene_to_obj_mapping, obj_
     return tokenizer, model, data_loader
 
 
+@spaces.GPU
 def get_model_response(model, tokenizer, data_loader, scene_id, user_input, max_new_tokens=50, temperature=0.2, top_p=0.9):
     input_data = [
         {