Yifan Liu

Results 2 comments of Yifan Liu

@bbbbubble @mengxuyiGit @RenieWell @SunzeY Hello, if the problem persists, maybe you can try my solution: change the cd_attention_mid attribute in File "./configs/train/stage1-mix-6views-lvis.yaml " from false to true, and change line...

Sorry, I was using Wonder3D's checkpoint as input for my 1st stage training, and that is fit. If you want to train from lambdalabs/sd-image-variations-diffusers, my method will not be working....