remove load in 8bit

app.py

@@ -44,7 +44,7 @@ def get_pipeline(model_name):
     #     chatbot = pipeline(model=model_name, task="text-generation", device=device)
 
     tokenizer = AutoTokenizer.from_pretrained(model_name, token=TOKEN)
-    model = AutoModelForCausalLM.from_pretrained(model_name, token=TOKEN, load_in_8bit=True)
+    model = AutoModelForCausalLM.from_pretrained(model_name, token=TOKEN)
     # chatbot = pipeline("conversational", model=model, tokenizer=tokenizer, device=device)
     chatbot = pipeline("conversational", model=model, tokenizer=tokenizer)
     return chatbot