Update app_v4.py

app_v4.py

@@ -33,7 +33,7 @@ if device == "cuda:0":
 tokenizer = AutoTokenizer.from_pretrained(pretrained_model_dir, use_fast=False)
 
 # Attempt to load the model, catch any OOM errors
-@st.cache_data
+@st.cache_resource
 def load_gptq_model():
     model = AutoGPTQForCausalLM.from_quantized(
         pretrained_model_dir,
@@ -45,10 +45,8 @@ def load_gptq_model():
     return model
     
 # Attempt to load the model, catch any OOM errors
-model_loaded = False
 try:
     model = load_gptq_model()
-    model_loaded = True
 except RuntimeError as e:
     if 'CUDA out of memory' in str(e):
         st.error("CUDA out of memory while loading the model. Try reducing the model size or restarting the app.")