After clicking “Watch Now” you will be prompted to login or join.
MLPerf: A Benchmark Suite for Machine Learning from an Academic-Industry Cooperative
David Kanter, Independent
GTC 2020
The ML field requires systematic benchmarking that represents real-world uses and enables fair comparisons across different software and hardware platforms. MLPerf is a machine-learning benchmark standard, and suite, driven by the industry (50+ companies) and researchers (800+) at large. The suite comprises a set of key machine-learning training and inference workloads representing important production-use cases, ranging from image classification and object detection to recommendation. MLPerf has released two rounds of training results and a new inference benchmark suite.