Update app.py

app.py

@@ -14,12 +14,16 @@ else:
 # Load the model and config when the script starts
 peft_model_id = "phearion/bigbrain-v0.0.1"
 config = PeftConfig.from_pretrained(peft_model_id)
-model = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path, return_dict=True)
+model = AutoModelForCausalLM.from_pretrained(
+    config.base_model_name_or_path,                          
+    low_cpu_mem_usage=True,
+    return_dict=True,
+    torch_dtype=torch.bfloat16)
 tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)
 
 # Load the Lora model
 model = PeftModel.from_pretrained(model, peft_model_id)
-
+model = model.merge_and_unload()
 
 def greet(text):
     batch = tokenizer(f"\"{text}\" ->: ", return_tensors='pt')