-
-
Notifications
You must be signed in to change notification settings - Fork 441
New issue
Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? # to your account
自定义模型 训练 保存 再倒入使用报错 #463
Comments
2.0.2 版本应该修复了该问题,可以测试一下~ |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. |
仍然报错 |
我也遇到了类似的问题,我的解决方法是将自定义的模型放到kashgari.tasks.classification 中,重新导入模型解决了此问题。不知有没有别的load_model()方法. |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. |
File ".\getvec.py", line 86, in
pro_model =Double_BiLSTM_Model.load_model('double_model')
File "d:\anaconda3\lib\site-packages\kashgari\tasks\abs_task_model.py", line 101, in load_model
model = load_data_object(model_config)
File "d:\anaconda3\lib\site-packages\kashgari\utils\serialize.py", line 25, in load_data_object
obj: Any = pydoc.locate(module_name)(**data['config'], **kwargs) # type: ignore
TypeError: 'NoneType' object is not callable
The text was updated successfully, but these errors were encountered: