Turkish Base Bert Uncased Mean Nli Stsb Tr

🚀 土耳其無大小寫區分基礎BERT模型用於平均NLI和STSB任務（土耳其語）

這是一個句子轉換器模型，它能將句子和段落映射到768維的密集向量空間，可用於聚類或語義搜索等任務。

該模型改編自ytu-ce-cosmos/turkish-base-bert-uncased，並在以下數據集上進行了微調：

• nli_tr
• emrecan/stsb-mt-turkish

⚠️ 重要提示

如模型作者所述，所有文本都需手動轉換為小寫：

text.replace("I", "ı").lower()

🚀 快速開始

📦 安裝指南

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）

from sentence_transformers import SentenceTransformer
sentences = ["Bu örnek bir cümle", "Her cümle dönüştürülür"]

model = SentenceTransformer('atasoglu/turkish-base-bert-uncased-mean-nli-stsb-tr')
embeddings = model.encode(sentences)
print(embeddings)

高級用法（使用HuggingFace Transformers庫）

不使用sentence-transformers時，可按以下步驟使用該模型：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 = ["Bu örnek bir cümle", "Her cümle dönüştürülür"]

# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('atasoglu/turkish-base-bert-uncased-mean-nli-stsb-tr')
model = AutoModel.from_pretrained('atasoglu/turkish-base-bert-uncased-mean-nli-stsb-tr')

# 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)

📚 詳細文檔

評估結果

在STS-b測試集上取得的結果如下：

Cosine-Similarity :	    Pearson: 0.8401	Spearman: 0.8410
Manhattan-Distance:	    Pearson: 0.8256	Spearman: 0.8261
Euclidean-Distance:	    Pearson: 0.8261	Spearman: 0.8268
Dot-Product-Similarity:	Pearson: 0.7823	Spearman: 0.7723

訓練參數

該模型的訓練參數如下：

數據加載器

torch.utils.data.dataloader.DataLoader，長度為90，參數如下：

{'batch_size': 64, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}

損失函數

sentence_transformers.losses.CosineSimilarityLoss.CosineSimilarityLoss

fit()方法的參數：

{
    "epochs": 4,
    "evaluation_steps": 9,
    "evaluator": "sentence_transformers.evaluation.EmbeddingSimilarityEvaluator.EmbeddingSimilarityEvaluator",
    "max_grad_norm": 1,
    "optimizer_class": "<class 'torch.optim.adamw.AdamW'>",
    "optimizer_params": {
        "lr": 2e-05
    },
    "scheduler": "WarmupLinear",
    "steps_per_epoch": null,
    "warmup_steps": 36,
    "weight_decay": 0.01
}

完整模型架構

SentenceTransformer(
  (0): Transformer({'max_seq_length': 256, '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})
)

📄 許可證

本項目採用MIT許可證。