Recent Posts

Manas Sahni, Shreya Varshini, Alind Khare, Alexey Tumanov || Systems for AI Lab (SAIL), Georgia Institute of Technology
Apr 28, 2021 5 min read

CompOFA improves the speed, cost, and usability of jointly training models for many deployment targets. By highlighting insights on model design and system deployment, we try to address an important problem for real-world usability of DNNs.
Aug 26, 2019 16 min read

It’s no surprise that modern deep-learning libraries have production-level, highly-optimized implementations of most operations. But just what is the black magic that these libraries use that we mere mortals don’t? What exactly does one do to “optimize” or accelerate neural networks operations?
Last updated on Dec 7, 2018 13 min read

Deploying efficient neural nets on mobiles is becoming increasingly important. This post explores the concept of quantized inference, and how it works in TensorFlow Lite.