Completed
Is there a type of model where it works particularly well and the type where it doesn't?
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
How to Make Your CPU as Fast as a GPU - Advances in Sparsity with Nir Shavit
Automatically move to the next video in the Classroom when playback concludes
- 1 Introduction
- 2 Sponsor: AssemblyAI
- 3 Start of Interview
- 4 How the NIR company was founded?
- 5 What is Sparsity about?
- 6 Link between the human brain and sparsity
- 7 Where should the extra resource that the human brain doesn't have go?
- 8 Analogy for Sparse Architecture
- 9 Possible future for Sparse Architecture as standard architure for Neural Networks
- 10 Pruning & Sparsification
- 11 What keeps us from building sparse models?
- 12 Why are GPUs so unsuited for sparse models?
- 13 CPU and GPU in connection with memory
- 14 What Neural Magic does?
- 15 How do you deal with overlaps in tensor columns?
- 16 The best type of sparsity to execute tons of CPU
- 17 What kind of architecture would make the best use out of a combined system of CPUs and GPUs?
- 18 Graph Neural Networks in connection to sparsity
- 19 Intrinsic connection between the Sparsification of Neural Networks, Non Layer-Wise Computation, Blockchain Technology, Smart Contracts and Distributed Computing
- 20 Neural Magic's target audience
- 21 Is there a type of model where it works particularly well and the type where it doesn't?
