Downloads · 30 days
89
5% of all-time downloads
aodiniz/bert_uncased_L-10_H-512_A-8_cord19-200616_squad2
bert_uncased_L-10_H-512_A-8_cord19-200616_squad2 is a question answering model from aodiniz. Use it when the input is a question plus a passage. It is set up for transformers.
BERT model with 10 Transformer layers and hidden embedding of size 512, referenced in Well-Read Students Learn Better: On the Importance of Pre-training Compact Models, fine-tuned for MLM on CORD-19 dataset (as releas…
Downloads · 30 days
89
5% of all-time downloads
All-time downloads
1.9K
Public
Repo size
629 MB
Likes
0
Public
Click a slice to open those files.
.bin191 MB · 44%
From the Hugging Face model README
BERT model with 10 Transformer layers and hidden embedding of size 512, referenced in Well-Read Students Learn Better: On the Importance of Pre-training Compact Models, fine-tuned for MLM on CORD-19 dataset (as released on 2020/06/16) and fine-tuned for QA on SQuAD v2.0.
python run_squad.py
--model_type bert
--model_name_or_path aodiniz/bert_uncased_L-10_H-512_A-8_cord19-200616
--train_file 'train-v2.0.json'
--predict_file 'dev-v2.0.json'
--do_train
--do_eval
--do_lower_case
--version_2_with_negative
--max_seq_length 384
--per_gpu_train_batch_size 10
--learning_rate 3e-5
--num_train_epochs 2
--output_dir bert_uncased_L-10_H-512_A-8_cord19-200616_squad2