Rasa version : 1.10.7
Rasa SDK version (if used & relevant): 1.10.7
Rasa X version (if used & relevant): 1.10.7
Python version : 3.7
Operating system (windows, osx, …): windows
Issue : In the latest rasa as above, I still get the “violation access” error. I have seen the related issues but they don’t help. I also found that when my training data is small like 7KB, it works fine without the error, but when my training data is larger like 642KB, the error appears, the config files are exactly same. And the time the error appear is just when the program is trying to save the model in windows user’s temp directory. So I doubt if there is anything to do with temp directory e.g. the model is too large to store in temp directory?
Error (including full traceback) :
Part II: train segment classifier now do training num training samples: 17267 2020-07-29 12:47:34 INFO rasa.nlu.model - Finished training component. Traceback (most recent call last): File "E:\PycharmProjects\rasa-1.10.7\rasa\train.py", line 453, in _train_nlu_async persist_nlu_training_data=persist_nlu_training_data, File "E:\PycharmProjects\rasa-1.10.7\rasa\train.py", line 482, in _train_nlu_with_validated_data persist_nlu_training_data=persist_nlu_training_data, File "E:\PycharmProjects\rasa-1.10.7\rasa\nlu\train.py", line 94, in train path, persistor, fixed_model_name, persist_nlu_training_data File "E:\PycharmProjects\rasa-1.10.7\rasa\nlu\model.py", line 231, in persist update = component.persist(file_name, dir_name) File "E:\PycharmProjects\rasa-1.10.7\rasa\nlu\extractors\mitie_entity_extractor.py", line 174, in persist self.ner.save_to_disk(entity_extractor_file, pure_model=True) File "D:\Anaconda3\envs\py37\lib\site-packages\mitie\mitie.py", line 271, in save_to_disk if (_f.mitie_save_named_entity_extractor_pure_model(filename, self.__obj) != 0): OSError: exception: access violation reading 0x0000000000000030 Process finished with exit code 1