Bert2bert Shared Turkish Summarization

🚀 Turkish BERT2BERT (shared) fine-tuned on MLSUM TR for summarization

This project presents a Turkish BERT2BERT model fine - tuned on the MLSUM TR dataset for text summarization. It offers an effective solution for summarizing Turkish news articles.

📚 Documentation

Model

The base model used is dbmdz/bert-base-turkish-cased (BERT Checkpoint).

Dataset

MLSUM is the first large - scale MultiLingual SUMmarization dataset. It is sourced from online newspapers and contains over 1.5M article/summary pairs in five languages: French, German, Spanish, Russian, and Turkish. When combined with English newspapers from the popular CNN/Daily mail dataset, it forms a large - scale multilingual dataset that can inspire new research in the text summarization field. Cross - lingual comparative analyses based on state - of - the - art systems have been reported, highlighting existing biases that justify the use of a multilingual dataset.

MLSUM tu/tr

Results

💻 Usage Examples

Basic Usage

import torch
from transformers import BertTokenizerFast, EncoderDecoderModel
device = 'cuda' if torch.cuda.is_available() else 'cpu'
ckpt = 'mrm8488/bert2bert_shared-turkish-summarization'
tokenizer = BertTokenizerFast.from_pretrained(ckpt)
model = EncoderDecoderModel.from_pretrained(ckpt).to(device)

def generate_summary(text):

    inputs = tokenizer([text], padding="max_length", truncation=True, max_length=512, return_tensors="pt")
    input_ids = inputs.input_ids.to(device)
    attention_mask = inputs.attention_mask.to(device)
    output = model.generate(input_ids, attention_mask=attention_mask)
    return tokenizer.decode(output[0], skip_special_tokens=True)
    
text = "Your text here..."
generate_summary(text)

Created by Manuel Romero/@mrm8488 with the support of Narrativa

Made with ♥ in Spain