Benchmarking Efficiency
Sara and Daniel discuss the challenges of benchmarking efficiency in machine learning, highlighting the limitations of MLPeRF and hardware agnostic metrics. They explore the need for a more comprehensive approach to benchmarking that captures real-world efficiency gains and address the incentives of benchmarking for hardware vendors.In this clip
From this podcast

The Gradient
Sara Hooker: Cohere For AI, the Hardware Lottery, and DL Tradeoffs
Related Questions