Code for running the random, herding, k-center baselines

Question

Code for running the random, herding, k-center baselines

GooLiang opened this issue a year ago · 4 comments

Hi! thanks for your awesome work.
I am trying to reproduce the results of random, herding, k-center baseline results in the paper.
As mentioned in the paper, using the implementation provided by "DATASET CONDENSATION WITH GRADIENT MATCHING" for Herding, K-Center, and DC.
However, "DATASET CONDENSATION WITH GRADIENT MATCHING"'s code also does not provide source code implementation.
Could you share the source code used to run the random, herding, k-center algorithm?

Answer 1 · 2023-12-13T02:12:03.000Z

Hi we have provided the code for these in train_coreset.py.

Answer 2 · 2023-12-13T02:17:00.000Z

Appreciate for your help!

Answer 3 · 2024-01-17T09:42:17.000Z

Hello author, can you provide the source code of DC-Graph?

Answer 4 · 2024-01-24T13:44:55.000Z

You can make simple adjustment on GCond to get DC-Graph:

replace the adjacency matrix A of the original with a zero matrix (which removes all the existing edges).
fix the condensed adjacency matrix A' to an identity matrix.