问题
I'm befuddled as to how best fix this. In the training data, it looks like the Google agent parsed out the information correctly.
But when I get it in my endpoint to fulfill the request, they do not match.
There is no way for me to indicate that it made a mistake because according to the training information, it parsed everything correctly. Has anyone else encountered this, and if so, what was the solution?
回答1:
Okay, I think I'm beginning to understand this a little. When this occurs, this means the format that the training data is recognizing is not actually in the sample utterances of that intent. That is why the data is mismatched. All I had to do was approve that training data.
来源：https://stackoverflow.com/questions/49500817/google-assistant-training-parse-does-not-match-data