Swedish BERT Fine-tuned on SQuAD v2
This model is a fine-tuning checkpoint of Swedish BERT on SQuAD v2.
Training data
Fine-tuning was done based on the pre-trained model KB/bert-base-swedish-cased.
Training and dev datasets are our Swedish translation of SQuAD v2.
Here is the HuggingFace Datasets.
Hyperparameters
batch_size = 16
n_epochs = 2
max_seq_len = 386
learning_rate = 3e-5
warmup_steps = 2900 # warmup_proportion = 0.2
doc_stride=128
max_query_length=64
Eval results
'exact': 66.72642524202223
'f1': 70.11149581003404
'total': 11156
'HasAns_exact': 55.574745730186144
'HasAns_f1': 62.821693965983044
'HasAns_total': 5211
'NoAns_exact': 76.50126156433979
'NoAns_f1': 76.50126156433979
'NoAns_total': 5945
Limitations and bias
This model may contain biases due to mistranslations of the SQuAD dataset.
BibTeX entry and citation info
@misc{svSQuADbert,
author = {Susumu Okazawa},
title = {Swedish BERT Fine-tuned on Swedish SQuAD 2.0},
year = {2021},
howpublished = {\url{https://huggingface.co/susumu2357/bert-base-swedish-squad2}},
}
- Downloads last month
- 376
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.