--------
Overview
--------

This folder contains the code used to generate results in J on CIFAR10 in the paper. Due to space limitations in the supplementary material, the embedding used in the paper is not provided. However, the code includes functionality for generating an appropriate embedding.

--------------------------------------------
Instructions to Reproduce Results on CIFAR10
--------------------------------------------

Set different N, M, R, P, \tau(fixed trajectory) as needed.

* CIFAR10 data
    Run: to get the CIFAR10 data
        python CIFAR-RES50_data.py

* SMC_parallel
    Run: to generate raw data  
        python CIFAR-RES50_psmc_hmc.py seed numwork  
    Set seed=1,2,...,RP manually. The numwork parameter corresponds to the number of CPU cores allocated per node on HPC.  
    Run: to compute full metrics  
        CIFAR_results_PSMC.m

* HMC_parallel
    Run: to generate raw data  
        python CIFAR-RES50_phmc.py seed
    Set seed=1,2,...,RP manually. 
    Run: to compute full metrics  
        CIFAR10_results_PHMC.m
