all agent alignment api update architecture benchmark business cuda dataset deployment embedding fine tuning hype inference infrastructure library ner new model open source optimization plugin prompt engineering quantization rag research security testing tool tutorial visualization workflow
Llm.c – LLM training in simple, pure C/CUDA
HN AI Stories · 779d ago · 9 · open source library inference tutorial

llm.c is a high-performance C/CUDA implementation for LLM pretraining that eliminates heavy dependencies (PyTorch, Python) while achieving 7% faster performance than PyTorch Nightly. It provides clean reference implementations for reproducing GPT-2/GPT-3 models with both GPU (CUDA) and CPU code paths, making it valuable for understanding model training mechanics and CUDA optimization.

Llamafile lets you distribute and run LLMs with a single file
HN AI Stories · 910d ago · 8 · tool open source deployment inference

llamafile 0.10.0 update from Mozilla.ai enables distributing and running open LLMs as single-file executables across platforms with no installation required, now with improved alignment to latest llama.cpp versions and support for more recent models. The tool also includes whisperfile for single-file speech-to-text capabilities, making local LLM deployment significantly more accessible for developers.