This is the 218 to 77 token distilled model, with vision.
#1
by Felldude - opened
This is the 218 to 77 token distilled model, with vision.
Felldude changed discussion title from I have failed to finetune the 218 CLIP-G with exisiting Vision for use in HiDream to This is the 218 to 77 token distilled model, with vision.