modules/llm_model.py

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline
from langchain.llms import HuggingFacePipeline

def load_llm_pipeline():
    model_id = "declare-lab/flan-alpaca-large"  # Better formatting
    tokenizer = AutoTokenizer.from_pretrained(model_id)
    model = AutoModelForSeq2SeqLM.from_pretrained(model_id)
    pipe = pipeline(
        "text2text-generation",
        model=model,
        tokenizer=tokenizer,
        max_length=512,
        do_sample=True,
        temperature=0.5,
    )
    return HuggingFacePipeline(pipeline=pipe)