local-llm-server/llm_server/llm/__init__.py

from llm_server.llm import oobabooga, vllm
from llm_server.routes.cache import redis


def get_token_count(prompt: str):
    backend_mode = redis.get('backend_mode', str)
    if backend_mode == 'vllm':
        return vllm.tokenize(prompt)
    elif backend_mode == 'ooba':
        return oobabooga.tokenize(prompt)
    else:
        raise Exception(backend_mode)