dudoxu's picture
Update README.md
7988b20
|
raw
history blame
636 Bytes
metadata
language: de
license: mit
tags:
  - german

We fine-tuned a German Question Answering model based on electra-base-german-uncased with our own German Question Answering dataset (deQuAD) containing 130k training and 11k test QA pairs.

Overview

  • Language model: electra-base-german-uncased
  • Language: German
  • Training data: deQuAD2.0 training set (~42MB)
  • Evaluation data: deQuAD2.0 test set (~4MB)
  • Infrastructure: 8xV100 GPU

Evaluation

We benchmarked the question answering performance on our deQuAD test data.