NVIDIA NIM

NVIDIA free reference containers to self-host optimized inference microservices.