CNTK Python API-继续训练模型 [英] CNTK python api - continue training a model

问题描述

我尝试了示例

该示例运行正常.我可以重新开始训练，但是在训练模型后无法继续训练.我尝试过

The example works fine. I can restart the training, but am unable to continue the training after a model has been trained. I have tried

agent.brain.model.load_model('setBdqn.mod', False) 

但收到错误消息.

推荐答案

错误消息指出load_model函数没有2个参数.您需要将其更改为

The error message says that the load_model function does not have 2 arguments. You need to change that to

agent.brain.model.load_model('setBdqn.mod')

我假设您将对save_model的调用更改为agent.brain.model.save_model('setBdqn.mod', False)-本教程示例将模型保存到名为dqn.mod的文件中.

I'm assuming that you changed the call to save_model to be agent.brain.model.save_model('setBdqn.mod', False) - the tutorial examples saves the model to a file named dqn.mod.

这篇关于CNTK Python API-继续训练模型的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！