Update README.md
Browse files
README.md
CHANGED
@@ -22,7 +22,7 @@ models.
|
|
22 |
|
23 |
# Model Description
|
24 |
|
25 |
-
|
26 |
|
27 |
Huggingface library doesn't implement Layer-Wise decay feature, which affects the performance on SQuAD task. The reported result of BioM-ELECTRA-Base-SQuAD in our paper is 84.4 (F1) since we use ELECTRA open-source code with TF checkpoint, which uses Layer-Wise decay. You can downoad our TensorFlow checkpoint that was fine-tuned on SQuAD2.0 and achieved 84.4 F1 score from here https://github.com/salrowili/BioM-Transformers .
|
28 |
|
22 |
|
23 |
# Model Description
|
24 |
|
25 |
+
We fine-tuned BioM-ELECTRA-Base, which was pre-trained on PubMed Abstracts, on the SQuAD2.0 dataset. Fine-tuning the biomedical language model on the SQuAD dataset helps improve the score on the BioASQ challenge. If you plan to work with BioASQ or biomedical QA tasks, it's better to use this model over BioM-ELECTRA-Base.
|
26 |
|
27 |
Huggingface library doesn't implement Layer-Wise decay feature, which affects the performance on SQuAD task. The reported result of BioM-ELECTRA-Base-SQuAD in our paper is 84.4 (F1) since we use ELECTRA open-source code with TF checkpoint, which uses Layer-Wise decay. You can downoad our TensorFlow checkpoint that was fine-tuned on SQuAD2.0 and achieved 84.4 F1 score from here https://github.com/salrowili/BioM-Transformers .
|
28 |
|