Open shersoni610 opened 1 year ago
It is mentioned in this blog: https://ai.meta.com/blog/dino-v2-computer-vision-self-supervised-learning In summary, SAM focuses on segmentation task, with some nice feature like promptable capability and DINOv2 can do segmentation task and also other tasks such as depth estimation, similarity search and so on.
Hello,
very much confused by these models. Both are claimed to be groundbreaking, so not sure on the strength and weakness of each method on segmentation tasks in particular.
Also, clip vs dinov2 is another confusing area. What are some advantages/disadvantages of each methods?