Hamming ball + DMALA update + Ising and Potts Update

[KatayoonGoshvadi]

Hi Hanjun, Please find below the following changes.

• Hamming Ball algorithm with hamming dist of one. Results match with what Haoran has (tested on categorical models).
• Debugged DMALA algorithm (the tests on Ising model were showing incorrect results, source of issue was shape related which got fixed).
• There was slight discrepancy between Ising and Potts results. Updated the models to match exactly with Haoran's code.
• Updated run configs for bernoulli, ising, potts and categorical model. They now include all the sampler and model variations.