app.py

from transformers import AutoTokenizer, AutoModelForCausalLM

model_id = "meta-llama/Meta-Llama-Guard-2-8B"

tokenizer = AutoTokenizer.from_pretrained(model_id, use_auth_token=True)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", torch_dtype="auto")

def ask_llama(prompt):
    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
    output = model.generate(**inputs, max_new_tokens=200)
    response = tokenizer.decode(output[0], skip_special_tokens=True)
    return response