Hierarchical SLM inference on NVIDIA Jetson AGX Orin. Runs Qwen2.5-3B via llama.cpp with CUDA acceleration, served over Tailscale VPN to an Orin NX wireless client.
websocket slm nvidia-jetson edge-ai tailscale llm jetson-orin jetson-agx-orin llama-cpp qwen2-5 hislm
-
Updated
Jun 25, 2026 - Python