- Deep CNN with AlexNet for Classifying/Recognizing objects
- Built using Python, TensorFlow, TFlearn and AWS EC2
- Used g2.8x GPU instances to speed up the classification process
- Built a model with an error rate of 6.2%
- Used factional max pooling with sparse CNN to improve accuracy
Technical Specifications
- Python 2.7
- TensorFlow 0.10.0
- TFLearn 0.2.1
- CUDA 8
- CuDnn v5
- g2.8x
- 80GB SSD
- 32 vCPU
sudo apt-get update
sudo apt-get -y dist-upgrade
sudo apt-get install python
sudo apt install python3-pip
sudo apt-get install -y libglu1-mesa libxi-dev libxmu-dev libglu1-mesa-dev gcc g++ gfortran build-essential git wget linux-image-generic libopenblas-dev python-dev liblapack-dev libblas-dev build-essential cmake git unzip pkg-config linux-image-generic linux-image-extra-virtual linux-source linux-headers-generic
sudo apt-get install zlib1g-dev python-imaging
export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.10.0rc0-cp27-none-linux_x86_64.whl
sudo pip install --upgrade $TF_BINARY_URL
sudo apt-get install python
sudo apt install python-pip
sudo pip install --upgrade pip
wget -qO- https://github.com/tflearn/tflearn/tarball/0.2.1 | tar xvz
cd tflearn-tflearn-a55c1fd/
sudo python setup.py install
sudo pip install pillow numpy scipy h5py
https://developer.nvidia.com/cuda-downloads
https://developer.nvidia.com/cudnn
python alex_net.py