Legal-BERTimbau-sts-large

Uso (Sentence-Transformers)

pip install -U sentence-transformers

from sentence_transformers import SentenceTransformer
sentences = ["Isto é um exemplo", "Isto é um outro exemplo"]

model = SentenceTransformer('rufimelo/Legal-BERTimbau-sts-large')
embeddings = model.encode(sentences)
print(embeddings)

Uso (HuggingFace Transformers)

from transformers import AutoTokenizer, AutoModel
import torch

# Mean Pooling - Tomar en cuenta la atención para un promedio correcto
def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0] # El primer elemento de model_output contiene todos los embeddings de los 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)

# Oraciones para las que queremos embeddings de oraciones
sentences = ['This is an example sentence', 'Each sentence is converted']

# Cargar modelo desde HuggingFace Hub
# Tokenize sentences
tokenizer = AutoTokenizer.from_pretrained('rufimelo/Legal-BERTimbau-sts-large')
model = AutoModel.from_pretrained('rufimelo/Legal-BERTimbau-sts-large')

# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')

# Computar embeddings de tokens
with torch.no_grad():
    model_output = model(**encoded_input)

# Realizar pooling. En este caso, pooling promedio.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)