Language model: gelectra-base-germanquad-distilled
Training data: GermanQuAD train set (~ 12MB)
Eval data: GermanQuAD test set (~ 5MB)
Infrastructure: 1x V100 GPU
Published: Apr 21st, 2021
- We trained a German question answering model with a gelectra-base model as its basis.
- The dataset is GermanQuAD, a new, German language dataset, which we hand-annotated and published online.
- The training dataset is one-way annotated and contains 11518 questions and 11518 answers, while the test dataset is three-way annotated so that there are 2204 questions and with 2204·3−76 = 6536answers, because we removed 76 wrong answers.
- In addition to the annotations in GermanQuAD, haystack's distillation feature was used for training. deepset/gelectra-large-germanquad was used as the teacher model.
See https://deepset.ai/germanquad for more details and dataset download in SQuAD format.
batch_size = 24 n_epochs = 6 max_seq_len = 384 learning_rate = 3e-5 lr_schedule = LinearWarmup embeds_dropout_prob = 0.1 temperature = 2 distillation_loss_weight = 0.75
We evaluated the extractive question answering performance on our GermanQuAD test set. Model types and training data are included in the model name. For finetuning XLM-Roberta, we use the English SQuAD v2.0 dataset. The GELECTRA models are warm started on the German translation of SQuAD v1.1 and finetuned on \\germanquad. The human baseline was computed for the 3-way test set by taking one answer as prediction and the other two as ground truth.
"exact": 62.4773139745916 "f1": 80.9488017070188
- Timo Möller:
timo.moeller [at] deepset.ai
- Julian Risch:
julian.risch [at] deepset.ai
- Malte Pietsch:
malte.pietsch [at] deepset.ai
- Michel Bartels:
michel.bartels [at] deepset.ai
Our focus: Industry specific language models & large scale QA systems.
Some of our work:
- German BERT (aka "bert-base-german-cased")
- GermanQuAD and GermanDPR datasets and models (aka "gelectra-base-germanquad", "gbert-base-germandpr")
By the way: we're hiring!
- Downloads last month