Completed
Obstacles of Sparsity Optimization
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
SparTA - Deep-Learning Model Sparsity via Tensor-with-Sparsity-Attribute
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 Computation Capacity vs DNN Model Size
- 3 Sparsity Commonly Exists
- 4 Evolving of Sparsity Pattern
- 5 Obstacles of Sparsity Optimization
- 6 The Myth of Proxy Metrics
- 7 Across-Stack Innovations in Silos
- 8 SparTA: An End-to-End Approach to Model Sparsity
- 9 Core Abstraction: TeSA
- 10 System Architecture
- 11 Execution Transformation
- 12 Code Specialization
- 13 What SparTA Achieves
- 14 Evaluation on Various Patterns & Models
- 15 End-to-end Opportunity
- 16 Mixed Sparsity Evaluation
- 17 Real Latency for Algorithm
- 18 Conclusion
