Hugging Face's logo
Turhan Can Kargın's picture
🤝 Open to Collab

Turhan Can Kargın

turhancan97
2 13 25
osamashabih's profile picture anilkask's profile picture hmb's profile picture
·
https://turhancankargin.me/

AI & ML interests

Computer Vision, Deep Learning, Machine Learning, Robotics, Reinforcement Learning, Control Theory

Recent Activity

posted an update about 2 hours ago
New blog post: Seeing ViT Registers 🔍 I reproduced the main observation of "Vision Transformers Need Registers" (Darcet et al.) on six DINOv2 checkpoints from the Hub: Base, Large and Giant, each with and without registers. Three findings: 1️⃣ Model size matters. In DINOv2 Giant, ~1.6% of patch tokens form a separate high-norm group. In Large the effect is weak, and in Base I found none. 2️⃣ Where you measure matters. The artifact tokens have very high norms before the final LayerNorm (hidden_states[-1]), but the lowest norms after it (last_hidden_state). 3️⃣ Registers take over. With registers, patch maps are smooth, and one register token carries a norm of ~1,346 instead. Everything is inference-only and reproducible with one command. The post also includes practical tips for people who use DINOv2 features for dense tasks or probing. 📝 Blog: https://huggingface.co/blog/turhancan97/seeing-vit-registers 💻 Code: https://github.com/turhancan97/seeing-vit-registers
upvoted an article about 2 hours ago
Seeing ViT Registers: Model Size, LayerNorm, and Where the High-Norm Tokens Go
published an article about 2 hours ago
Seeing ViT Registers: Model Size, LayerNorm, and Where the High-Norm Tokens Go
View all activity

Organizations

Bohemian Visual Recognition Alliance's profile picture Keras Dreambooth Event's profile picture scikit-learn's profile picture Blog-explorers's profile picture