sts-GBERT-bi-encoder

Uso (Sentence-Transformers)

Usar este modelo es fácil cuando tienes sentence-transformers instalado:

pip install -U sentence-transformers

Luego puedes usar el modelo de esta manera:

from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('sts-GBERT-bi-encoder')
embeddings = model.encode(sentences)
print(embeddings)

Uso (HuggingFace Transformers)

Sin sentence-transformers, puedes usar el modelo de esta manera: Primero, pasas tu entrada través del modelo transformer, luego debes aplicar la operación de pooling correcta sobre las embeddings de palabras contextualizadas.

from transformers import AutoTokenizer, AutoModel
import torch
# Mean Pooling - Tomar en cuenta la máscara de atención para un promedio correcto
def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0]  # El primer elemento de model_output contiene todas las embeddings de tokens
    input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
    return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)

# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('sts-GBERT-bi-encoder')
model = AutoModel.from_pretrained('sts-GBERT-bi-encoder')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
    model_output = model(**encoded_input)
# Perform pooling. In this case, max pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)