Spaces:
Sleeping
Sleeping
File size: 1,011 Bytes
8a0c27f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 |
from typing import Sequence
import numpy as np
from sentence_transformers import SentenceTransformer
class Vectorizer(object):
"""
TODO
"""
def __init__(self, model_name: str = 'all-MiniLM-L6-v2') -> None:
"""
Initialize the vectorizer with a pre-trained embedding model.
"""
self.model = SentenceTransformer(model_name)
def transform(self, prompts: Sequence[str]) -> np.ndarray:
"""
Transform texts into numerical vectors using the specified model.
"""
return self.model.encode(list(prompts))
@staticmethod
def cosine_similarity(query_vector: np.ndarray, corpus_vectors: np.ndarray) -> np.ndarray:
"""
Calculate cosine similarity between prompt vectors.
"""
query_norm = query_vector / np.linalg.norm(query_vector)
corpus_norms = corpus_vectors / np.linalg.norm(corpus_vectors, axis=1, keepdims=True)
return np.dot(corpus_norms, query_norm.T).flatten()
|