NeuML/bert-small-cord19

BERT-Small fine-tuned on CORD-19 dataset

BERT L6_H-512_A-8 model fine-tuned on the CORD-19 dataset.

CORD-19 data subset

The training data for this dataset is stored as a Kaggle dataset. The training data is a subset of the full corpus, focusing on high-quality, study-design detected articles.

Building the model

python run_language_modeling.py
    --model_type bert
    --model_name_or_path google/bert_uncased_L-6_H-512_A-8
    --do_train
    --mlm
    --line_by_line
    --block_size 512
    --train_data_file cord19.txt
    --per_gpu_train_batch_size 4
    --learning_rate 3e-5
    --num_train_epochs 3.0
    --output_dir bert-small-cord19
    --save_steps 0
    --overwrite_output_dir