NVTabular is a feature engineering and preprocessing library for tabular data designed to quickly and easily manipulate terabyte scale datasets used to train deep learning based recommender systems.
I benchmarked NVTabular+TensorFlow for the criteo example and numbers do not as expected. I think we need a follow-up / profile investigation in NVTabular latency numbers.
What questions are you trying to answer? Please describe. Analyze latency/throughput for NVTabular+TensorFlow for Triton Inference Server