deepset
/

gelectra-large-germanquad

Question Answering

Inference Endpoints

Model card Files Files and versions Community

Tuana commited on Jul 19, 2022

Commit

1b7c5a7

•

1 Parent(s): 475addd

Update README.md

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -38,9 +38,9 @@ embeds_dropout_prob = 0.1
 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.
-![performancetable](https://lh3.google.com/u/0/d/1IFqkq8OZ7TFnGzxmW6eoxXSYa12f2M7O=w1970-h1546-iv1)
 ## Authors
  **Timo Möller:** [email protected]

 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](https://deepset.ai/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.
+![performancetable](https://images.prismic.io/deepset/1c63afd8-40e6-4fd9-85c4-0dbb81996183_german-qa-vs-xlm-r.png)
 ## Authors
  **Timo Möller:** [email protected]