utils/customLLM.py

from typing import Any, List, Mapping, Optional

from langchain.llms.base import LLM
from transformers import Pipeline


class CustomLLM(LLM):
    pipeline = None
    # Create the pipeline for question answering
    def __init__(self, model_pipeline: Pipeline):
        self.pipeline = model_pipeline

    def _call(self, prompt: str, stop: Optional[List[str]] = None) -> str:
        prompt_length = len(prompt)
        response = self.pipeline(prompt, max_new_tokens=525)[0]["generated_text"]

        # only return newly generated tokens
        return response[prompt_length:]

    @property
    def _identifying_params(self) -> Mapping[str, Any]:
        return {"name_of_model": self.model_name}

    @property
    def _llm_type(self) -> str:
        return "custom"