GGUF
A single-file format for quantised models, used by llama.cpp and other tools to run models on ordinary computers.
Also: llama.cpp
Georgi Gerganov released llama.cpp in March 2023, a program in plain C and C++ that ran Meta's LLaMA on a laptop processor. Its model format evolved into GGUF in August 2023, which packs the weights, usually quantised to between 2 and 8 bits, the tokenizer and settings into one file. Tools such as Ollama and LM Studio build on it, and community uploaders publish GGUF versions of most open models within days of release. It made running models locally practical for millions of people.