This project contains over 100 schematic diagrams of large model technology principles, systematically introducing large models and reinforcement learning, covering LLM/VLM large model architectures, training algorithms (RL, RLHF, GRPO, DPO, SFT, CoT), performance optimization, and RAG, etc.