Index Translate for AI Pod
Select 2B, 9B or 35B-A3B-preview in the AI Pod model store. Jetson AGX Orin uses Orinfer; Jetson AGX Thor/T5000 uses vLLM. Model weights live in separate size/backend repositories. This repository contains the complete weight-free Docker-save runtimes, controller, source, licenses, performance evidence and minimal LPK. All artifacts are mirrored on ModelScope and Hugging Face with immutable revisions and complete SHA256 verification. Original models: https://github.com/bilibili/Index-Translate Orinfer: https://github.com/iMMIQ/orinfer (LGPL-3.0-or-later). Each dependency retains its supplied license.