Summary of files:
- /data directory includes the 4 datasets we used
- datasets.py contains functions for reading each dataset
- models.py contains the update steps for SGD, SAGA, SSVRG on Logistic Regression and Matrix Factorization models
- main.py runs the experiment necessary to produce the results shown in Figure 1. It takes three arguments, the dataset, arrival distribution, and processing rate rho/lambda. For these input parameters, it outputs a data file that can be used by plot-main.py.
- plot-main.py produces plots like those shown in Figure 1. It requires the necessary data files have been written by main.py first.
- sensitivity.py runs the experiment necessary to produce the results shown in Figure 2. It takes two arguments, the processing rate rho/lambda, and the amount of skew M/lambda.
- plot-sensitivity.py produces plots shown in Figure 2. It requires the necessary data files have been written by sensitivity.py first.
