harshitkhandelwal208/hk

Local LLMs that run lean, load instantly and keep getting smarter. A from-scratch runtime + model format in Zig: zero-copy mmap, CPU/CUDA/Vulkan engines, OpenAI-compatible server, 2:4 sparsity, Net2Net growth, in-file version history, self-learning loops.

语言：Zig · GitHub 星标：24 · Fork：3

话题：ai · cuda · gguf · huggingface · inference · llm · mmap · model-container

在 GitHub 上查看原仓库