GGUF에는 가중치 외에 무엇이 들어 있고, 아직 무엇이 빠져 있나?

GGUF serves as the standard file format for llama.cpp-based local LLM applications, consolidating model weights, chat templates, special tokens, and sampler configurations into a single file to significantly improve deployment efficiency and compatibility. While recent additions like Jinja2 template support and explicit sampler chain ordering have enhanced usability, critical standardization gaps remain—including tool calling format inconsistency, missing think_token support, and multimodal capability detection—revealing clear directions for future ecosystem improvements.