FICHA · AUR
ik-llama.cpp-cuda
llama.cpp fork with additional SOTA quants and improved performance (CUDA Backend)
- local-llm-inference-tool
- CLI
- TERMINAL
- AI
- Launchable
- Runs in terminal
· official+codex · reviewed · Jun 1, 2026 description in en
Local large-language-model inference can use NVIDIA CUDA acceleration with extra quantization formats.
It is useful for users running offline AI models on compatible GPUs. Local models can still process sensitive text and may generate wrong output, while GPU builds require matching drivers and CUDA support.
How to run
llama-cli
Commands: llama-cli
Permissions
Permissions not analysed for this source yet.