How new libraries can improve model performance with OctoML's Luis Ceze

"I always have this mystery where a new version of cuDNN comes out, and the models get way faster with just a better library."

Luis Ceze, co-founder and CEO of OctoML, explains how new libraries can improve performance on operations like convolutions and matrix multiplication.

---

This clip is part of "OctoML CEO Luis Ceze on accelerating machine learning systems", an episode of the Gradient Dissent podcast.

🌟 Full episode here:    • Luis Ceze — Accelerating Machine Lear...   🌟
Show notes (transcription & links) 👉 http://wandb.me/gd-luis-ceze

---

Check out Fully Connected, which features curated machine learning reports by researchers exploring deep learning techniques, Kagglers showcasing winning models, industry leaders sharing best practices, and more:
https://wandb.ai/fully-connected