DeepCCA
fork from (https://github.com/shamidreza/deepcca)
A python code for deep canonical correlation analysis (dcca) written using numpy
Requirements:
1- numpy
2- theano
Datasets:
-
Mnist halves:
Built from data/mnist.pkl.gz (link)
-
Mnist same digit
Built in cca_linear.py
dA.py : denoising autoencoder
sdAE.py : Stacking denoising AE
mlp.py : standard Multi-layer perceptron
my_sda.py : Modified deepCCA (that actually works!)