Moco Sentencedistilbertv2.1

🚀 moco-sentencedistilbertV2.1

This is a sentence embedding model that maps sentences and paragraphs into a 768-dimensional dense vector space, which can be used for tasks such as clustering or semantic search.

Pipeline Tag

sentence-similarity

Tags

• sentence-transformers
• feature-extraction
• sentence-similarity
• transformers
• ko
• en

Widget

• Source Sentence: "대한민국의 수도는?"
• Sentences:
  • "서울특별시는 한국이 정치,경제,문화 중심 도시이다."
  • "부산은 대한민국의 제2의 도시이자 최대의 해양 물류 도시이다."
  • "제주도는 대한민국에서 유명한 관광지이다"
  • "Seoul is the capital of Korea"
  • "울산광역시는 대한민국 남동부 해안에 있는 광역시이다"

🚀 Quick Start

This is a sentence-transformers model. It maps sentences and paragraphs to a 768-dimensional dense vector space and can be used for tasks like clustering or semantic search.

• This model is created by converting the bongsoo/mdistilbertV2.1 MLM model into a SentenceBERT model and then performing additional STS teacher-student distillation training.
• Vocab: 152,537 (32,989 new vocabs are added to the original 119,548 vocabs).

📦 Installation

Using Sentence-Transformers

pip install -U sentence_transformers

Using HuggingFace Transformers

pip install transformers[torch]

💻 Usage Examples

Basic Usage (Sentence-Transformers)

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)

# Calculate cosine_scores using sklearn
# => The input embeddings should be 2D, like (1,768).
from sklearn.metrics.pairwise import paired_cosine_distances, paired_euclidean_distances, paired_manhattan_distances
cosine_scores = 1 - (paired_cosine_distances(embeddings[0].reshape(1,-1), embeddings[1].reshape(1,-1)))

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

Outputs

[[ 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

Advanced Usage (HuggingFace Transformers)

Without sentence-transformers, you can use the model as follows: First, pass your input through the transformer model, then apply the appropriate pooling operation on top of the contextualized word embeddings.

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 = ["서울은 한국이 수도이다", "The capital of Korea is Seoul"]

# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('bongsoo/moco-sentencedistilbertV2.1')
model = AutoModel.from_pretrained('bongsoo/moco-sentencedistilbertV2.1')

# 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)

# Calculate cosine_scores using sklearn
# => The input embeddings should be 2D, like (1,768).
from sklearn.metrics.pairwise import paired_cosine_distances, paired_euclidean_distances, paired_manhattan_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]}')

Outputs

Sentence embeddings:
tensor([[ 0.2712, -0.5837,  0.0074,  ..., -0.0038,  0.0180, -0.0965],
        [ 0.2765, -0.5754,  0.0079,  ...,  0.0766, -0.0783, -0.0612]])
*cosine_score:0.9513546228408813

📚 Documentation

Evaluation Results

• For performance measurement, the following Korean (kor) and English (en) evaluation corpora are used:
  • Korean: korsts (1,379 sentence pairs) and klue-sts (519 sentence pairs)
  • English: stsb_multi_mt (1,376 sentence pairs) and glue:stsb (1,500 sentence pairs)
• The performance indicator is cosin.spearman/max (the maximum value among cosine, Euclidean, Manhattan, and doc).
• Refer to the evaluation measurement code here.

For an automated evaluation of this model, see the Sentence Embeddings Benchmark: https://seb.sbert.net

Training

The model was trained with the following parameters:

1. MLM Training

• Input Model: distilbert-base-multilingual-cased
• Corpus: Training - bongsoo/moco-corpus-kowiki2022 (7.6M), Evaluation - bongsoo/bongevalsmall
• Hyperparameters: Learning Rate: 5e-5, Epochs: 8, Batch Size: 32, Max Token Length: 128
• Vocab: 152,537 (32,989 new vocabs are added to the original 119,548 vocabs)
• Output Model: mdistilbertV2.1 (Size: 643MB)
• Training Time: 63h/1GPU (24GB/23.9GB used)
• Evaluation: Training Loss: 2.203400, Evaluation Loss: 2.972835, Perplexity: 23.43 (bong_eval: 1,500)
• Refer to the training code here.

2. STS Training

• Convert BERT to SentenceBERT.
• Input Model: mdistilbertV2.1 (Size: 643MB)
• Corpus: korsts (5,749) + kluestsV1.1 (11,668) + stsb_multi_mt (5,749) + mteb/sickr-sts (9,927) + glue stsb (5,749) (Total: 38,842)
• Hyperparameters: Learning Rate: 3e-5, Epochs: 800, Batch Size: 128, Max Token Length: 256
• Output Model: sbert-mdistilbertV2.1 (Size: 640MB)
• Training Time: 13h/1GPU (24GB/16.1GB used)
• Evaluation (cosin_spearman): 0.790 (Corpus: korsts(tune_test.tsv))
• Refer to the training code here.

3. Distillation Training

• Student Model: sbert-mdistilbertV2.1
• Teacher Model: paraphrase-multilingual-mpnet-base-v2 (max_token_len: 128)
• Corpus: news_talk_en_ko_train.tsv (English-Korean dialogue-news parallel corpus: 1.38M)
• Hyperparameters: Learning Rate: 5e-5, Epochs: 40, Batch Size: 128, Max Token Length: 128 (to match the teacher model)
• Output Model: sbert-mdistilbertV2.1-distil
• Training Time: 17h/1GPU (24GB/9GB used)
• Refer to the training code here.

4. STS Training

• Input Model: sbert-mdistilbertV2.1-distil
• Corpus: korsts (5,749) + kluestsV1.1 (11,668) + stsb_multi_mt (5,749) + mteb/sickr-sts (9,927) + glue stsb (5,749) (Total: 38,842)
• Hyperparameters: Learning Rate: 3e-5, Epochs: 1200, Batch Size: 128, Max Token Length: 256
• Output Model: moco-sentencedistilbertV2.1
• Training Time: 12h/1GPU (24GB/16.1GB used)
• Evaluation (cosin_spearman): 0.839 (Corpus: korsts(tune_test.tsv))
• Refer to the training code here.

For more details about the model creation process, refer to here.

Config

{
  "_name_or_path": "../../data11/model/sbert/sbert-mdistilbertV2.1-distil",
  "activation": "gelu",
  "architectures": [
    "DistilBertModel"
  ],
  "attention_dropout": 0.1,
  "dim": 768,
  "dropout": 0.1,
  "hidden_dim": 3072,
  "initializer_range": 0.02,
  "max_position_embeddings": 512,
  "model_type": "distilbert",
  "n_heads": 12,
  "n_layers": 6,
  "output_past": true,
  "pad_token_id": 0,
  "qa_dropout": 0.1,
  "seq_classif_dropout": 0.2,
  "sinusoidal_pos_embds": false,
  "tie_weights_": true,
  "torch_dtype": "float32",
  "transformers_version": "4.21.2",
  "vocab_size": 152537
}

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 256, 'do_lower_case': False}) with Transformer model: DistilBertModel 
  (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})
)

Tokenizer Config

{
  "cls_token": "[CLS]",
  "do_basic_tokenize": true,
  "do_lower_case": false,
  "mask_token": "[MASK]",
  "max_len": 128,
  "name_or_path": "../../data11/model/sbert/sbert-mdistilbertV2.1-distil",
  "never_split": null,
  "pad_token": "[PAD]",
  "sep_token": "[SEP]",
  "special_tokens_map_file": "../../data11/model/distilbert/mdistilbertV2.1-4/special_tokens_map.json",
  "strip_accents": false,
  "tokenize_chinese_chars": true,
  "tokenizer_class": "DistilBertTokenizer",
  "unk_token": "[UNK]"
}

Sentence Bert Config

{
  "max_seq_length": 256,
  "do_lower_case": false
}

Config Sentence Transformers

{
  "__version__": {
    "sentence_transformers": "2.2.0",
    "transformers": "4.21.2",
    "pytorch": "1.10.1"
  }
}

📄 License

No license information provided.

Citing & Authors

bongsoo