models/llama2.py

from transformers import AutoTokenizer, AutoModelForCausalLM

model_name = "meta-llama/Llama-2-13b-chat-hf"

def load():
  global model
  global tokenizer
  
  model = AutoModelForCausalLM.from_pretrained(model_name)
  tokenizer = AutoTokenizer.from_pretrained(model_name)

def generate(input_text):
  input_ids = tokenizer.encode(input_text, return_tensors="pt")
  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=200, num_beams=2, eos_token_id=tokenizer.eos_token_id)
  
  return tokenizer.decode(output_ids[0], skip_special_tokens=True)