German BERT2BERT fine-tuned on MLSUM DE for summarization
Model
bert-base-german-cased (BERT Checkpoint)
Dataset
MLSUM is the first large-scale MultiLingual SUMmarization dataset. Obtained from online newspapers, it contains 1.5M+ article/summary pairs in five different languages -- namely, French, German, Spanish, Russian, Turkish. Together with English newspapers from the popular CNN/Daily mail dataset, the collected data form a large scale multilingual dataset which can enable new research directions for the text summarization community. We report cross-lingual comparative analyses based on state-of-the-art systems. These highlight existing biases which motivate the use of a multi-lingual dataset. MLSUM de
Results
| Set | Metric | # Score |
|---|---|---|
| Test | Rouge2 - mid -precision | 33.04 |
| Test | Rouge2 - mid - recall | 33.83 |
| Test | Rouge2 - mid - fmeasure | 33.15 |
Usage
import torch
from transformers import BertTokenizerFast, EncoderDecoderModel
device = 'cuda' if torch.cuda.is_available() else 'cpu'
ckpt = 'mrm8488/bert2bert_shared-german-finetuned-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