9x Model Serving Performance Without Changing…

Martynas Šubonis

Nov 4, 2024

Achieve Up to 9x Faster & 13x Smaller Model Serving Compared to Naive Setups

2 Comments

Great job, thanks!

Have you benchmarked this implementation on NVIDIA GPUs?

Martynas Šubonis

Hey Bogdan, glad you enjoyed this. Unfortunately, I never got around to benchmarking this on NVIDIA hardware

#nojs-banner { position: fixed; bottom: 0; left: 0; padding: 16px 16px 16px 32px; width: 100%; box-sizing: border-box; background: red; color: white; font-family: -apple-system, "Segoe UI", Roboto, Helvetica, Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol"; font-size: 13px; line-height: 13px; } #nojs-banner a { color: inherit; text-decoration: underline; } This site requires JavaScript to run correctly. Please turn on JavaScript or unblock scripts