Why is my deepeval running slowly?

[QiQi-OvO]

from typing import Callable

from deepeval.models import DeepEvalBaseLLM


class ModelAdapter(DeepEvalBaseLLM):
    def __init__(self, generate_output_func: Callable, model_name: str):
        self.generate_func = generate_output_func
        self.model_name = model_name

    def get_model_name(self, *args, **kwargs) -> str:
        return self.model_name

    def load_model(self, *args, **kwargs):
        return self

    def generate(self, prompt: str) -> str:
        try:
            output = self.generate_func(prompt)
        except Exception as e:
            raise TypeError("generate func need 1 input prompt and 1 output string:", e)
        return output

    async def a_generate(self, prompt: str) -> str:
        return self.generate(prompt)

I used the code above to create a custom llm model, but it was very slow to import the package.
What should I do, except for the import phase, the actual running speed is also very slow. :(

[penguine-ip]

Because your a_generate() method is not actually async, try giving it an async implementation and run again @QiQi-OvO