handler.py

from typing import Dict, List, Any
from transformers import pipeline

import torch.nn.functional as F
from torch import Tensor
from transformers import AutoTokenizer, AutoModel

def average_pool(last_hidden_states: Tensor,
                 attention_mask: Tensor) -> Tensor:
    last_hidden = last_hidden_states.masked_fill(~attention_mask[..., None].bool(), 0.0)
    return last_hidden.sum(dim=1) / attention_mask.sum(dim=1)[..., None]

class EndpointHandler():
    def __init__(self, path=""):
        self.pipeline = pipeline("feature-extraction", model=path)
        self.tokenizer = AutoTokenizer.from_pretrained(path)
        self.model = AutoModel.from_pretrained(path)

    def __call__(self, data: Dict[str, Any]) -> List[List[int]]:
        inputs = data.pop("inputs",data)

        batch_dict = self.tokenizer(inputs, max_length=512, padding=True, truncation=True, return_tensors='pt')

        outputs = self.model(**batch_dict)

        embeddings = average_pool(outputs.last_hidden_state, batch_dict['attention_mask'])
        embeddings = F.normalize(embeddings, p=2, dim=1).tolist()

        return embeddings