This is a fork of the official MLPerf™ Inference v4.0 repository with changes to seamlessly run on Taiga Cloud machines.

The changes are made directly inside NVIDIA's submission as closest to our platform.

This repository is not meant to be or become an official benchmark submission.

Following the MLPerf Results Messaging Guidelines, all our results are not verified by MLCommons Association and when presenting we must accompany the numbers with a required text or footnote.