Update app.py

app.py

@@ -6,7 +6,7 @@ from transformers import LlamaForCausalLM, PreTrainedTokenizerFast, pipeline
 model_name = "Azurro/APT3-1B-Instruct-v1"
 
 tokenizer = PreTrainedTokenizerFast.from_pretrained(model_name)
-model = LlamaForCausalLM.from_pretrained(model_name, torch_dtype=torch.float16)
+model = LlamaForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16)
 
 def generate_text(prompt, max_length, temperature, top_k, top_p):
     prompt = f'<s>[INST] {prompt.strip()} [/INST]'