qudabear
qudabear
Good luck. Ben never answers a straight question. As for my own experiments, 1e-6 and 2e-6 only work for faces, and only with 10-20 images max, text encoder around 350-450....
This has been an issue from day one that Ben keeps claiming is resolved by adding the port to the server, but that fix does not work. It is present...
Incorrect. Resuming training with new images absolutely can and often does improve a model, but you should not re-train the text encoder. When training, you can remove problematic images and...
So who is going to refund me for the shitty code, the broken disconnects, the terrible training results, and now the random disconnect of the service I was using?
I highly suggest everyone on this thread put pressure on Runpod to restore the service, or demand refunds. I also highly suggest Ben improves his code and provides detailed settings...
> > I highly suggest everyone on this thread put pressure on Runpod to restore the service, or demand refunds. I also highly suggest Ben improves his code and provides...
> > > > > > > > > > > Ben's (Runpod) notebook has been broken or subpar for a long time and it never got fixed. > >...
> > > > > > > > > > > > > > > > > > > Ben's (Runpod) notebook has been broken or subpar for a long...