DINOv3 continual self-supervised pre-training. Domain-adapts public DINOv3 ViT backbones on unlabeled images via teacher-student self-distillation (DINO + iBOT + KoLeo, optional Gram anchoring) and converts the EMA teacher into a timm-format backbone for downstream tasks.
安装
npx skills add NVIDIA/skills --skill tao-train-dinov3npx skills add NVIDIA/skills --skill tao-train-dinov3 -a claude-code -gnpx skills add NVIDIA/skills --skill tao-train-dinov3 -a codex -ggemini skills install https://github.com/NVIDIA/skills.git --path skills/tao-train-dinov3gh skill install NVIDIA/skills tao-train-dinov3git clone --depth 1 https://github.com/NVIDIA/skills /tmp/NVIDIA__skills
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/NVIDIA__skills/skills/tao-train-dinov3 ~/.claude/skills/tao-train-dinov3 # Claude Code
cp -R /tmp/NVIDIA__skills/skills/tao-train-dinov3 ~/.agents/skills/tao-train-dinov3 # Codex, Cursor, Gemini CLI, Copilot, OpenCode