Pytorch in Production
Adrien discusses the evolution of using Pytorch in production, highlighting the initial challenges and subsequent advancements in deployment tools like Onyx and Tensorrt. The shift to Pytorch early on allowed for faster iteration in research, optimizing for the bottleneck of experimentation. Now, with improved tools, deploying models is more efficient and aligns with engineering standards.In this clip
From this podcast
Gradient Dissent - A Machine Learning Podcast
Adrien Gaidon — Advancing ML Research in Autonomous Vehicles
Related Questions