Showcase: Deep learning in the cloud. Building a handwritten character recognition model with GNU R and MXNet with Amazon Web Services on GPU instances.
Business story behind the showcase
Optical character recognition (OCR) for handwritten text is applicable in areas where several documents are being used in business for processing large amounts of paper documents. One example could be a traditional post operator that wants to automatically process information on the envelopes. The envelopes are usually not neatly adressed and standard OCR solutions fail in this area. A custom-made character recognition engine designed for this task might receive much better quality level.
Approach
The showcase presents two main things:
- Construction of deep laerning models with MXNet library
- How to run GPU computing in the cloud with Amazon Web Services
Prerequisites
We use bitfusion.io Scientific Computing AMI
The AMI contains Ubuntu 14 along with a R installation along with CUDA drivers. Additionally we have installed MXNet running the following commands:
sudo apt-get updatesudo apt-get install -y build-essential git libblas-dev libopencv-devgit clone --recursive https://github.com/dmlc/mxnet
Next modify config.mk by setting the following keys:
USE_CUDA = 1
USE_CUDA_PATH = /usr/local/cuda
USE_BLAS = atlas
Finally, compile mxnet with the command make –j4
Usage instruction
- Install R packages
01_install_packages.R - Prepare the dataset
02_download_datasets.R - Declare layers for the deep neural network
03_declare_mlp_model.R - Create data iterators for seuential data reading
04_prepare_data_iterators.R - Fit the model to the data
05_fit_mlp_model.R - If the fitting process is interrupted a script for resuming computation state can be used:
06_restart_mlp_model.R - Perform predictions and observe the results
07_predict_mlp_model.R
What next
This example presents one possible usage of deep learning models for classification of images. One important problem is selection of an optimal structure for a deep neural network. This requires execution of several experiments for measuring predictive capabilities for various network topologies. Amazon Web Services comes forward to this need and offers very large GPU instances. The flagship offering is a p2.16xlarge offering 16 x GPU Nvidia TESLA K80 with a total of 80'000 GPU cores. This machine availabe from around $2.10 on AWS spot market would make it to the list of Top Supercomputers just 10 years ago.