pipeline.py

from typing import List
from sentence_transformers import SentenceTransformer
import os

class PreTrainedPipeline:
    def __init__(self, path=""):
        """
        Initialize model
        """
        self.model = SentenceTransformer(os.path.join(path))
            # os.path.join(path, 'quora-distilbert-multilingual')
           #"sentence-transformers/quora-distilbert-multilingual"
        #)




    def __call__(self, inputs: str) -> List[float]:
        """
        Args:
            inputs (:obj:`str`):
                a string to get the features of.
        Return:
            A :obj:`list` of floats: The features computed by the model.
        """
        return self.model.encode(inputs).tolist()

#
if __name__ == "__main__":
    xx = PreTrainedPipeline()

    print(xx.__call__("hei"))