MLPerf Endpoints Session at GTC on March 19th at noon

[David Kanter]

Hello Everyone,

Good afternoon! We are inviting you to come to our session at GTC to learn about the next-generation of MLPerf benchmarking – and to find out how you can contribute to building the next industry-standard for performance measurement.

As you may know, we are building MLPerf Endpoints, a benchmark for Gen AI that is designed to be radically easier to use and more understandable to a broad audience (and importantly, your customers!).

Crucially, it measures the performance of API endpoints – how the vast majority of people interact with AI systems – and supports rolling submissions, so that your organization can submit and publish results at any point in time (e.g. for a conference, launch event, etc.), instead of the current 6 month cycle.

We believe the increased clarity, speed and flexibility of MLPerf Endpoints is essential to the entire ecosystem going forward.

We will be giving a talk and demonstration at GTC, including our visualization tools on March 19th at noon pacific time (https://www.nvidia.com/gtc/session-catalog/sessions/gtc26-s82276/). This is a team effort with participation all across the industry – NVIDIA, AMD, Intel, Google, and Krai have all included test systems in what we will preview and many other organizations contributed to defining and building the benchmarks. 

We expect that many of our contributors will attend this session. We would like to invite you and your colleagues to join us at this session to learn more and engage with the community.

Thank you all for your hard work, passion, and commitment to advancing machine learning. We look forward to seeing you at GTC.


Thanks,


David Kanter
MLCommons
Founder, Head of MLPerf