Open krrr8s opened 1 month ago
Hi @krrr8s, we're currently still in the company review process. We anticipate being able to release it after our paper is accepted.
Will it be possible to fine-tune DINOv2-G model (1.3B) with customed synthetic dataset, then assign pseudo depth labels on unlabeled real images, and then train a smaller model?
Hello! This work is very impressive and shows great potential. May I ask when the Depth-Anything-V2-Giant 1.3B model will be released? I am very eager for its release!