You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.

test_word_seg.py 979 B

12345678910111213141516171819202122232425262728
  1. from fastNLP.action.tester import Tester
  2. from fastNLP.action.trainer import WordSegTrainer
  3. from fastNLP.loader.base_loader import BaseLoader
  4. from fastNLP.models.word_seg_model import WordSeg
  5. def test_wordseg():
  6. train_config = WordSegTrainer.TrainConfig(epochs=5, validate=False, save_when_better=False,
  7. log_per_step=10, log_validation=False, batch_size=254)
  8. trainer = WordSegTrainer(train_config)
  9. model = WordSeg(100, 2, 1000)
  10. train_data = BaseLoader("load_train", "./data_for_tests/cws_train").load_lines()
  11. trainer.train(model, train_data)
  12. test_config = Tester.TestConfig(save_output=False, validate_in_training=False,
  13. save_dev_input=False, save_loss=False, batch_size=254)
  14. tester = Tester(test_config)
  15. test_data = BaseLoader("load_test", "./data_for_tests/cws_test").load_lines()
  16. tester.test(model, test_data)
  17. if __name__ == "__main__":
  18. test_wordseg()

一款轻量级的自然语言处理(NLP)工具包,目标是减少用户项目中的工程型代码,例如数据处理循环、训练循环、多卡运行等