train_examples, val_examples = load_train_val_examples(args)
trainer.train(args, train_examples, val_examples)
def do_eval(args):
args.model_path = args.best_model_path
eval_examples = load_eval_examples(eval_text_file, eval_bio_file)
model = load_model(args)
trainer.evaluate(args, model, eval_examples)
def do_predict(args):
args.model_path = args.best_model_path
test_examples = load_test_examples(args)
model = load_model(args)
trainer.predict(args, model, test_examples)
#do_eval(args)
#do_predict(args)
#print(trainer.pred_results)
if __name__ == '__main__':
def add_special_args(parser):
parser.add_argument("--generate_submission",
action="store_true",
help="")
return parser
args = get_args(experiment_params=experiment_params,
special_args=[add_special_args])
logger.info(f"args: {args}")
main(args)
args.train_file, test_data_generator,
args.test_file)
elif args.do_train:
train_examples, val_examples = load_train_val_examples(args)
trainer.train(args, train_examples, val_examples)
elif args.do_eval:
_, eval_examples = load_train_val_examples(args)
model = load_model(args)
trainer.evaluate(args, model, eval_examples)
elif args.do_predict:
test_examples = load_test_examples(args)
model = load_model(args)
trainer.predict(args, model, test_examples)
save_ner_preds(args, trainer.pred_results, test_examples)
if __name__ == '__main__':
def add_special_args(parser):
parser.add_argument("--do_eda", action="store_true", help="")
parser.add_argument("--generate_submission",
action="store_true",
help="")
return parser
args = get_args([add_special_args])
main(args)