src/embeddings/embedder.py

from transformers import AutoTokenizer, AutoModel
import torch
import numpy as np
from typing import List

class Embedder:
    def __init__(self, model_name: str = "BAAI/bge-m3"):
        self.tokenizer = AutoTokenizer.from_pretrained(model_name)
        self.model = AutoModel.from_pretrained(model_name)

    def embed(self, texts: List[str]) -> np.ndarray:
        inputs = self.tokenizer(texts, padding=True, truncation=True, return_tensors="pt")
        with torch.no_grad():
            outputs = self.model(**inputs)
        embeddings = outputs.last_hidden_state[:, 0]  # lấy embedding từ CLS token
        embeddings = torch.nn.functional.normalize(embeddings, p=2, dim=1)
        return embeddings.cpu().numpy()