From ddeecf08e67dc894bf82a7da24cea0ea8f61981e Mon Sep 17 00:00:00 2001 From: smanjil Date: Fri, 6 Nov 2020 09:21:13 +0100 Subject: [PATCH] german medbert model details (#8266) * model details * Apply suggestions from code review Co-authored-by: Julien Chaumond --- model_cards/smanjil/German-MedBERT/README.md | 33 ++++++++++++++++++++ 1 file changed, 33 insertions(+) create mode 100644 model_cards/smanjil/German-MedBERT/README.md diff --git a/model_cards/smanjil/German-MedBERT/README.md b/model_cards/smanjil/German-MedBERT/README.md new file mode 100644 index 0000000000..a1b86607d6 --- /dev/null +++ b/model_cards/smanjil/German-MedBERT/README.md @@ -0,0 +1,33 @@ +--- +language: de +--- + +# German Medical BERT + +This is a fine-tuned model on Medical domain for German language and based on German BERT. + +## Overview +**Language model:** bert-base-german-cased + +**Language:** German + +**Fine-tuning:** Medical articles (diseases, symptoms, therapies, etc..) + +**Eval data:** NTS-ICD-10 dataset (Classification) + +**Infrastructure:** Gogle Colab + + +## Details +- We fine-tuned using Pytorch with Huggingface library on Colab GPU. +- With standard parameter settings for fine-tuning as mentioned in original BERT's paper. +- Although had to train for upto 25 epochs for classification. + +## Performance (Micro precision, recall and f1 score for multilabel code classification) +![performance](https://raw.githubusercontent.com/smanjil/finetune-lm/master/performance.png) + +## Author +Manjil Shrestha: `shresthamanjil21 [at] gmail.com` + +Get in touch: +[LinkedIn](https://www.linkedin.com/in/manjil-shrestha-038527b4/)