We list the table of contents here.
.
    - ./Data 
	    - Jacobian_scaling: numpy array for APJN scaling with depth
        - Phase_diagram: numpy arrays or dictioneries for data used in phase diagrams.
        - Training: numpy arrays or dictioneries for training data. 

    - ./Notebooks
        - Demo.ipynb: a quick notebook to reproduce most of our plots, it takes around 20 min to run on a cpu.
	    - jacobian_fitting.ipynb: load data from './Data/Jacobian_scaling' to plot depth-scaling of APJN and find critical exponents.
        - phase_diagrams_and_training_results.ipynb: load data from './Data/Training' and './Data/Phase_diagram' to make phase diagrams and training curves.


    - ./utils
        - partialjaclib.py: our customized library, needed to run the 'Demo.ipynb' notebook. All models and functions we wrote are in this file (except ResNet ones).
	    - resnetlib.py: our customized library for training ResNet models.

    - Training_MLP_Fig3_Column1to3.py: learning rate search and training codes for first three columns of Fig 3.

    - Training_MLP_Fig3_Column4to6.py: learning rate search and training codes for last three columns of Fig 3.

    - Training_ResNet_Fig4.py: learning rate search and training codes for training curves in Fig 4 (3)(4)

    - Training_Mixer_Fig5: Mixer training file for Fig 5.
