Ko Sbert Multitask

🚀 ko-sbert-multitask

這是一個 sentence-transformers 模型，它可以將句子和段落映射到一個 768 維的密集向量空間，可用於聚類或語義搜索等任務。

🚀 快速開始

安裝依賴

若要使用此模型，需先安裝 sentence-transformers：

pip install -U sentence-transformers

使用示例

基礎用法

使用 sentence-transformers 庫調用該模型：

from sentence_transformers import SentenceTransformer
sentences = ["안녕하세요?", "한국어 문장 임베딩을 위한 버트 모델입니다."]

model = SentenceTransformer('jhgan/ko-sbert-multitask')
embeddings = model.encode(sentences)
print(embeddings)

高級用法

若不使用 sentence-transformers 庫，可按以下方式使用模型：首先將輸入傳遞給 Transformer 模型，然後對上下文詞嵌入應用正確的池化操作。

from transformers import AutoTokenizer, AutoModel
import torch


#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0] #First element of model_output contains all token embeddings
    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('jhgan/ko-sbert-multitask')
model = AutoModel.from_pretrained('jhgan/ko-sbert-multitask')

# 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, mean pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])

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

📊 評估結果

這是在 KorSTS、KorNLI 訓練數據集上進行多任務訓練後，使用 KorSTS 評估數據集進行評估的結果：

🔧 訓練細節

模型的訓練參數如下：

數據加載器 1

sentence_transformers.datasets.NoDuplicatesDataLoader.NoDuplicatesDataLoader，長度為 8885，參數如下：

{'batch_size': 64}

損失函數 1

sentence_transformers.losses.MultipleNegativesRankingLoss.MultipleNegativesRankingLoss，參數如下：

{'scale': 20.0, 'similarity_fct': 'cos_sim'}

數據加載器 2

torch.utils.data.dataloader.DataLoader，長度為 719，參數如下：

{'batch_size': 8, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}

損失函數 2

sentence_transformers.losses.CosineSimilarityLoss.CosineSimilarityLoss

訓練方法參數

{
    "epochs": 5,
    "evaluation_steps": 1000,
    "evaluator": "sentence_transformers.evaluation.EmbeddingSimilarityEvaluator.EmbeddingSimilarityEvaluator",
    "max_grad_norm": 1,
    "optimizer_class": "<class 'transformers.optimization.AdamW'>",
    "optimizer_params": {
        "lr": 2e-05
    },
    "scheduler": "WarmupLinear",
    "steps_per_epoch": null,
    "warmup_steps": 360,
    "weight_decay": 0.01
}

📚 完整模型架構

SentenceTransformer(
  (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)

📄 引用與作者

• Ham, J., Choe, Y. J., Park, K., Choi, I., & Soh, H. (2020). Kornli and korsts: New benchmark datasets for korean natural language understanding. arXiv preprint arXiv:2004.03289
• Reimers, Nils and Iryna Gurevych. “Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks.” ArXiv abs/1908.10084 (2019)
• Reimers, Nils and Iryna Gurevych. “Making Monolingual Sentence Embeddings Multilingual Using Knowledge Distillation.” EMNLP (2020).