zero123 Can we continue training on the basis of 105000.ckpt?

Can we continue training on the basis of 105000.ckpt?

Open TimenoLong opened this issue 1 year ago • 1 comments

May I ask if you can publish the yaml file of 105000.ckpt to us so that we can train on the model you have already trained?

Feb 24 '24 04:02 TimenoLong

i concur--is there a way to finetune models? Is there any chance z123 will ever support LoRa's or some similar?

Mar 02 '24 07:03 stvn037