search

IntelLabs / ais-benchmarks

A framework, based on python and numpy, for evaluation of sampling methods
MIT License
10 stars 3 forks source link

Implement visualization module #15

Open jfelip opened 1 year ago

jfelip commented 1 year ago

This module will show the evolution of sampling algorithms showing the ground truth and samples that are accepted/rejected, importance weights and so on.

The idea is to provide a video generation that can be used to explain how the method works and to debug that the method is behaving as expected.

jfelip commented 1 year ago

To implement this each sampling method should include a visualization flag. When the flag is on, the method should keep track of each sampling step. The tracked sampling steps will be used to show the animated sampling algorithm.