C inference engine; own loader, kernels, and tokenizer. Runs hako-sho (3B) and hako-koi (7B) on CPU. No llama.cpp, no ollama.
machine-learning tokenizer c99 simd transformer avx2 quantization no-dependencies inference-engine hako cpu-inference llm local-llm qwen llama-cpp-alternative mithraeum q4k ggml-alternative
-
Updated
Jul 26, 2026 - C