moco-sentencedistilbertV2.1

Uso (Sentence-Transformers)

Instalar sentence-transformers:

pip install -U sentence_transformers

Usar el modelo:

from sentence_transformers import SentenceTransformer
sentences = ["서울은 한국이 수도이다", "The capital of Korea is Seoul"]

model = SentenceTransformer('bongsoo/moco-sentencedistilbertV2.1')
embeddings = model.encode(sentences)
print(embeddings)

from sklearn.metrics.pairwise import paired_cosine_distances
cosine_scores = 1 - (paired_cosine_distances(embeddings[0].reshape(1,-1), embeddings[1].reshape(1,-1)))

print(f'*cosine_score:{cosine_scores[0]}')

Uso (HuggingFace Transformers)

Instalar transformers:

pip install transformers[torch]

Usar el modelo:

from transformers import AutoTokenizer, AutoModel
import torch

def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0]
    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 = ["서울은 한국이 수도이다", "The capital of Korea is Seoul"]

tokenizer = AutoTokenizer.from_pretrained('bongsoo/moco-sentencedistilbertV2.1')
model = AutoModel.from_pretrained('bongsoo/moco-sentencedistilbertV2.1')

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

with torch.no_grad():
    model_output = model(**encoded_input)

sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])

print("Sentence embeddings:")
print(sentence_embeddings)

from sklearn.metrics.pairwise import paired_cosine_distances
cosine_scores = 1 - (paired_cosine_distances(sentence_embeddings[0].reshape(1,-1), sentence_embeddings[1].reshape(1,-1)))

print(f'*cosine_score:{cosine_scores[0]}')

Resultados

# Embedded results
[[ 0.27124503 -0.5836643   0.00736023 ... -0.0038319   0.01802095 -0.09652182]
 [ 0.2765149  -0.5754248   0.00788184 ...  0.07659392 -0.07825544 -0.06120609]]
*cosine_score:0.9513546228408813

# Sentence embeddings results
Sentence embeddings:
censor([ 0.2712, -0.5837,  0.0074,  ..., -0.0038,  0.0180, -0.0965])
cosine_score:0.951354622848813