Streamlining the usage of huggingface run_language_model.py in finetuning or training from scratch a bert model with your own text data
python bert_downstream_classification.py --data_path=[data_path] --epoch=30 --batch_size=64 --min_each_group=3 --maxlength=30 --model_output=[model_path] --model_start --model_prediction
python bert_downstream_classification.py --data_path=[data_path_testing] --do_test --epoch=30 --batch_size=64 --min_each_group=3 --maxlength=30 --model_start=[model_path] --model_prediction=[prediction_path]