Title | ||
---|---|---|
Accelerating large-scale convolutional neural networks with parallel graphics multiprocessors |
Abstract | ||
---|---|---|
Training convolutional neural networks (CNNs) on large sets of high-resolution images is too computationally intense to be performed on commodity CPUs. Such architectures, however, achieve state-of-theart results on low-resolution machine vision tasks such as recognition of handwritten characters. We have adapted the inherent multi-level parallelism of CNNs for Nvidia's CUDA GPU architecture to accelerate the training by two orders of magnitude. This dramatic speedup permits to apply CNN architectures to pattern recognition tasks on datasets with high-resolution natural images. |
Year | DOI | Venue |
---|---|---|
2010 | 10.1007/978-3-642-15825-4_9 | ICANN (3) |
Keywords | Field | DocType |
handwritten character,dramatic speedup,cuda gpu architecture,high-resolution image,high-resolution natural image,cnn architecture,commodity cpus,training convolutional neural network,parallel graphics multiprocessors,inherent multi-level parallelism,pattern recognition task,large-scale convolutional neural network,machine vision,pattern recognition,high resolution,low resolution,neural network | Graphics,Machine vision,Shared memory,Convolutional neural network,CUDA,Computer science,Parallel computing,Artificial intelligence,Artificial neural network,Machine learning,Speedup | Conference |
Volume | ISSN | ISBN |
6354 | 0302-9743 | 3-642-15824-2 |
Citations | PageRank | References |
21 | 3.74 | 8 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Dominik Scherer | 1 | 243 | 22.67 |
Hannes Schulz | 2 | 55 | 10.82 |
Sven Behnke | 3 | 1672 | 181.84 |