Looking for inferincing vs training (and I saw this)

https://devblogs.nvidia.com/parallelforall/inference-next-step-gpu-accelerated-deep-learning/