Skip to content

Flash is a framework developed to test whether a Horn-sampler is epsilon-close or eta-far from a given distribution with confidence greater than 1-delta.

meelgroup/flash

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Flash - A Horn Sampler Tester

Flash is a framework developed to test whether a Horn-sampler is epsilon-close or eta-far from a given distribution with confidence greater than 1-delta.

The full paper has six Appendices(A-F).

Appendices A-E contain the theoretical justification for the theorems and lemmas mentioned in the paper.

Appendix F has the results for the entire set of benchmarks and all the tested samplers along with the comparisons of the performance of our testers with the baseline testers.

The results presented in the extended tables of Appendix F can be verified with the code and benchmarks presented in this folder.

Requirements to run the code

  • Python 3.6

To install the required libraries, run:


pip install -r requirements.txt

Getting Started

The directory is subdivided into 3 parts: uflash, flash, baseline.

  • The uflash directory contains codes and benchmarks to run uFlash. Also, one complete set of logs is already been given inside. To run uFlash you need to copy the sampler binaries inside the "codes" directory of uflash:

     cp -r samplers/ uflash/codes/
    
  • The flash directory contains codes and benchmarks to run Flash. Also, One complete set of logs is already been given inside. To run Flash you need to copy the sampler binaries inside the "codes" directory of flash:

     cp -r samplers/ flash/codes/
    
  • The baseine directory contains codes to obtain necessary information regarding baseline approach. To run the baseline codes again you need to copy the sampler binaries inside the directory:

     cp -r samplers/ baseline/
    

Contributors

About

Flash is a framework developed to test whether a Horn-sampler is epsilon-close or eta-far from a given distribution with confidence greater than 1-delta.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published