Even Faster CNNs: Exploring the New Class of Winograd Algorithms
Convolutional Neural Networks (CNNs) are compute-intensive, with increasingly complex architectures. Join us to discover how the new class of Winograd Algorithms can make CNNs faster than ever before, allowing workloads such as classification and recognition to be implemented on low-power, Arm-based platforms.
- The recently introduced class of algorithms that can reduce the arithmetic complexity of convolution layers with small filter sizes
- The latest optimization techniques for the most common solutions, such as GEMM
- The design of Winograd algorithms, with an analysis of the complexity and the performance achieved for convolutional neural networks