This post overviews the paper Confident Learning: Estimating Uncertainty in Dataset Labels authored by Curtis G. Northcutt, Lu Jiang, and Isaac L. Chuang.
In a previous post, Build a Pro Deep Learning Workstation… for Half the Price, I shared every detail to buy parts and build a professional quality deep learning rig for nearly half the cost of pre-built rigs from companies like...
The Benchmarking Keras PyTorch GitHub project benchmarks every pre-trained model in PyTorch and Keras (Tensorflow). All benchmarks are reproducible.
04/16/2019 Update: A better build is available in this post. 03/21/2019 Updates: Amazon links added for all parts. Added Blower-style GPU, faster/cheaper M.2 SSD, and other options.