Theano-based Large-Scale Visual Recognition with Multiple GPUs
arXiv:1412.2302
Abstract
In this report, we describe a Theano-based AlexNet (Krizhevsky et al., 2012) implementation and its naive data parallelism on multiple GPUs. Our performance on 2 GPUs is comparable with the state-of-art Caffe library (Jia et al., 2014) run on 1 GPU. To the best of our knowledge, this is the first open-source Python-based AlexNet implementation to-date.
ICLR 2015 workshop camera-ready version
References in corpus (6)
- Theano: new features and speed improvements
- cuDNN: Efficient Primitives for Deep Learning
- One weird trick for parallelizing convolutional neural networks
- PyCUDA and PyOpenCL: A Scripting-Based Approach to GPU Run-Time Code Generation
- Pylearn2: a machine learning research library
- GPU Asynchronous Stochastic Gradient Descent to Speed Up Neural Network Training
Cited by in corpus (9)
- Comparative Study of Deep Learning Software Frameworks
- Exploiting Local Structures with the Kronecker Layer in Convolutional Networks
- Exploring the Design Space of Deep Convolutional Neural Networks at Large Scale
- A detailed comparative study of open source deep learning frameworks
- Theano-MPI: a Theano-based Distributed Training Framework
- Improving Transparency of Deep Neural Inference Process
- Cost-Sensitive Deep Learning with Layer-Wise Cost Estimation
- Similarities and differences between stimulus tuning in the inferotemporal visual cortex and convolutional networks
- Network Analysis for Explanation