This Convolutional Neural Network (CNN) is built using TensorFlow to analyse images of the cosmic web to measure the equation of state paramter w.
The CNN is fed a number of images created from density maps of the dark matter distribution created from simulations with different values of w and then trained to identify the cosmological parameters of the N-body simulation.
Update: You can now specify the size the sizes of the images in Mpc/h but I haven't added rotation yet. The images don't have to be square.
An additional code to calculate the power spectrum is included as a sanity check for the density cubes - the linear power spectra derived from CAMB should match up with the power spectra measured directly from the cubes on large scales.