Infinigence AI Open-Sources APXInf for Embodied Edge Inference on Jetson Thor

Infinigence AI with Tsinghua and SJTU has open-sourced APXInf, an edge inference engine for embodied models on Jetson and desktop GPUs. On Jetson Thor, the PI 0.5 model in FP8 accelerated from ~278 ms to under 26 ms (about 38.46 Hz), roughly a 10x speedup.

• PI 0.5 FP8 latency on Jetson Thor cut from ~278 ms to <26 ms
• Supports Jetson Orin, Jetson Thor, and GeForce RTX 4090
• Rust runtime with Python bindings, repo RLinf/APXinf-robo
• Plans include nvfp4, AMD and domestic chip backends