search

Danield21 / Rationale4CDECR

MIT License
1 stars 0 forks source link

Some questions about the results #4

Open era211 opened 5 hours ago

era211 commented 5 hours ago

Hello author, when I was replicating the code, I found that the results of the first epoch were good, and since the second epoch, the results of each epoch have remained the same.May I ask why this is? image

Danield21 commented 4 hours ago

Hi, the results look definately very weird. I have uploaded all of my training log in outputs which are normal.
Did you strictly use the code and config in the repo without any modification? Could you please tell me the code and config you used for such result?

era211 commented 2 hours ago

Hi, the results look definately very weird. I have uploaded all of my training log in outputs which are normal. Did you strictly use the code and config in the repo without any modification? Could you please tell me the code and config you used for such result?

it's ok