MPDCompress - Matrix Permutation Decomposition Algorithm for Deep Neural Network Compression
This addresses the challenge of efficient DNN deployment on resource-constrained hardware, offering a method that balances compression and performance, though it appears incremental as it builds on existing compression techniques.
The paper tackles the problem of deploying large deep neural networks on edge devices by proposing MPDCompress, a compression algorithm that molds weight matrices into hardware-friendly block structures, achieving up to 10x compression with less than 1% accuracy loss on models like LeNet and AlexNet, and up to 4x inference speedup on mobile GPUs.
Deep neural networks (DNNs) have become the state-of-the-art technique for machine learning tasks in various applications. However, due to their size and the computational complexity, large DNNs are not readily deployable on edge devices in real-time. To manage complexity and accelerate computation, network compression techniques based on pruning and quantization have been proposed and shown to be effective in reducing network size. However, such network compression can result in irregular matrix structures that are mismatched with modern hardware-accelerated platforms, such as graphics processing units (GPUs) designed to perform the DNN matrix multiplications in a structured (block-based) way. We propose MPDCompress, a DNN compression algorithm based on matrix permutation decomposition via random mask generation. In-training application of the masks molds the synaptic weight connection matrix to a sub-graph separation format. Aided by the random permutations, a hardware-desirable block matrix is generated, allowing for a more efficient implementation and compression of the network. To show versatility, we empirically verify MPDCompress on several network models, compression rates, and image datasets. On the LeNet 300-100 model (MNIST dataset), Deep MNIST, and CIFAR10, we achieve 10 X network compression with less than 1% accuracy loss compared to non-compressed accuracy performance. On AlexNet for the full ImageNet ILSVRC-2012 dataset, we achieve 8 X network compression with less than 1% accuracy loss, with top-5 and top-1 accuracies of 79.6% and 56.4%, respectively. Finally, we observe that the algorithm can offer inference speedups across various hardware platforms, with 4 X faster operation achieved on several mobile GPUs.