Update app.py

app.py

@@ -50,7 +50,7 @@ if torch.cuda.is_available():
     model = AutoModelForCausalLM.from_pretrained(
         model_id, device_map="cuda", torch_dtype=torch.bfloat16
     )
-    reft_model = ReftModel.load("Ayushnangia/Lossfunk-Residency-Llama-3-8B-Instruct", model, from_huggingface_hub=True)
+    reft_model = ReftModel.load("Ayushnangia/Lossfunk-Residency-Llama-3-8B-Instruct", model, from_huggingface_hub=True, weights_only=True)
     reft_model.set_device("cuda")
     tokenizer = AutoTokenizer.from_pretrained(model_id)
     tokenizer.use_default_system_prompt = True