Update app.py

app.py

@@ -1,5 +1,5 @@
 import gradio as gr
-from transformers import AutoProcessor, PaliGemmaForConditionalGeneration, AutoModelForSeq2SeqLM
+from transformers import AutoProcessor, PaliGemmaForConditionalGeneration, BitsAndBytesConfig
 from peft import PeftModel
 import spaces
 import torch
@@ -8,13 +8,17 @@ import os
 token = os.getenv('token')
 HfFolder.save_token(token)
 device = "cuda"
-
-model = PaliGemmaForConditionalGeneration.from_pretrained("triphuong57/paligemma_lora").to(device)
+model = PaliGemmaForConditionalGeneration.from_pretrained("triphuong57/paligemma_lora")
 processor = AutoProcessor.from_pretrained("google/paligemma-3b-mix-224")
 
-@spaces.GPU(duration=180)
+@spaces.GPU(duration=240)
 def greet(image, prompt):
     
+    
+    # model = PaliGemmaForConditionalGeneration.from_pretrained("/folders", torch_dtype=torch.float16, quantization_config=quantization_config).to(device)
+    
+    
+    # # model = PeftModel(base_model, "/folders").to(device)
     model_inputs = processor(text=prompt, images=image, return_tensors="pt")
     input_len = model_inputs["input_ids"].shape[-1]
     with torch.inference_mode():