Tensorgrams and Deep Learning
Greg explains the origins of tensorgrams and how they relate to deep learning. He breaks down the complexity of deep learning into matrix multiplication and nonlinearities, allowing for a universal understanding of the fundamentals. The low-level representation of tensorgrams provides insights into network width and its impact on training.In this clip
From this podcast

The Gradient
Greg Yang on Communicating Research, Tensor Programs, and µTransfer
Related Questions