Collection of best practices, optimization guides, architecture overview, performance counters

FlashInfer: Kernel Library for LLM Serving

All-in-one training for vision models (YOLO, ViTs, RT-DETR, DINOv3): pretraining, fine-tuning, distillation.