timm DINOv3 Collection Meta AI's DINOv3 weights in timm. ViTs with `qkvb` have a zero QV bias present, otherwise bias is disabled. QKV bias are all 0 in original weights. • 18 items • Updated Sep 19, 2025 • 40
DETRPose: Real-time end-to-end transformer model for multi-person pose estimation Paper • 2506.13027 • Published Jun 16, 2025 • 2
LINEA: Fast and Accurate Line Detection Using Scalable Transformers Paper • 2505.16264 • Published May 22, 2025
ViTPose Collection Collection for ViTPose models based on transformers implementation. • 10 items • Updated Jan 12, 2025 • 18