issue4

[gejianqiao]

Hello esteemed professor, I have benefited greatly from your code, but I have a question for you. It's because I started open AI encoding, following your structure. However, after over 100 rounds of training, CSMSE didn't decrease; instead, it rose to over 700. Do you know which reason? but if you could answer me, I would be very grateful.

[bartam1]

Hello @gejianqiao! I tried to reproduce this experiment with dow30 but for me the results RANKIC and others was far behind what the paper said. Have you tried also to reproducde using this source code? Something is not good here.