Disclaimer

Performance Disclaimer

Benchmarks published through the Service are run by Metrum AI using specific hardware/software configurations, datasets, and model servers at the time of testing. Performance varies widely based on model selection and version, framework/library revisions, drivers/firmware, precision settings, data preprocessing, application complexity, the quantity of data, batch sizes, software versions, libraries used, and other factors. Benchmark results are frequently updated, and previously published results may become outdated. All results are provided for informational purposes only and are not a promise or guarantee of performance. Individual results may vary.

Metrum AI makes no warranties regarding the accuracy or completeness of performance data. Users should conduct their own testing on production-representative workloads and configurations before making purchasing or implementation decisions. For questions about performance testing methodology or to report issues, please contact our support team at insights@metrum.ai.