cmhungsteve/Awesome-Transformer-Attention

Organizes transformer and attention papers across 15+ vision tasks—from classification and detection to video analysis, 3D point clouds, medical imaging, and low-level restoration—with separate curated lists for multi-modal and emerging applications. Papers are categorized by architectural variant (pure attention, conv-stem hybrids, efficient transformers, attention-free alternatives) and continuously updated with recent conference proceedings from CVPR, ICCV, NeurIPS, and ICML. Includes implementation links and tutorial resources alongside paper references to support practitioners working with PyTorch/TensorFlow transformer frameworks across diverse vision domains.

5,022 stars. No commits in the last 6 months.