What DINO saw: ALiBi positional encoding reduces positional bias in Vision Transformers
Paper • 2603.16840 • Published
ALiBi models created by Pawlowsky, M. and Docherty, R. https://arxiv.org/abs/2603.16840
base DINO models by META
SAM models used from the onnx-community
Model tree for GandalfTheHolyOne/DinoSawGUI
Base model
facebook/dinov2-small