llama.cpp

94ed3d86 - scripts: add build-llamafile.sh for creating llamafile bundles

scripts: add build-llamafile.sh for creating llamafile bundles Add a script that fetches llamafile tools from Mozilla and builds self-contained executables bundling LLM models with inference code. Features: - Downloads llamafile tools (zipalign, base executables) from releases - Supports local GGUF models or downloading from URLs - Server mode option for web UI builds - Configurable cache directory and output paths - Embedded default arguments support - Retry logic with exponential backoff for downloads - Platform detection (Linux, macOS, Windows)