The result predicted by the TXT model is inconsistent with the prediction of xlearn itself

[Changzhisong]

I used XLearn to train the fm model and then parse the generated TXT model (according to the FM formula), but the result of my final prediction is inconsistent with the result of xlearn itself predict, the difference is quite large. After I checked the code I wrote many times, and I verified it with self-test data, there is no problem, so the problem is:

1. Did xlearn process the data after prediction? (I looked at the source code, it seems that I have not done it, I wonder if it is missing)
2. xlearn saves the TXT model if there is a problem;

[xulile]

I have a similar problem ! Do you know the reason，now?

[whk6688]

一样的疑问 很渴望指导 谢谢了