Update app_v4.py

app_v4.py

@@ -35,7 +35,7 @@ tokenizer = AutoTokenizer.from_pretrained(pretrained_model_dir, use_fast=False)
 # Attempt to load the model, catch any OOM errors
 @st.cache_data
 def load_gptq_model():
-    AutoGPTQForCausalLM.from_quantized(
+    model = AutoGPTQForCausalLM.from_quantized(
         pretrained_model_dir,
         model_basename="Jackson2-4bit-128g-GPTQ",
         use_safetensors=True,
@@ -44,6 +44,7 @@ def load_gptq_model():
     model.eval()  # Set the model to inference mode
     return model
     
+# Attempt to load the model, catch any OOM errors
 model_loaded = False
 try:
     model = load_gptq_model()