Results and documentation related to my 2-month stay at the SUnAI group (Universitat Oberta de Catalunya). The stay is related to (convolutional) Deep Neural Networks and the experiments are carried out using state-of-the-art tools such as Caffe (BVLC), TensorFlow (Google) and Torch.
Go to the experiments section to see the different tests and whacky experiments carried out with neural nets or visit the documentation and resources section for a list of interesting publications and tools (including some remarks and comments).
Remember to download the required image databases and models by running setup.sh before you run any of the experiments.
All the experiments require the CAFFE_ROOT environment variable to be set to the location of your caffe instalation. Also, adding $CAFFE_ROOT/build/tools to your PATH can be handy. For instance, you can place the following line in your ~/.bash_profile file:
export CAFFE_ROOT=$HOME/caffe
export PATH=$CAFFE_ROOT/build/tools:$PATH