Run the full 2.78-trillion-parameter Kimi K3 model beyond available RAM by streaming activated weights directly from NVMe. A dependency-free, embeddable C inference engine.
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).