How can Train KAN on batches of data

[lukmanulhakeem97]

While training KAN using entire samples say 10000, end up in memory overflow issue. Instead of giving entire data at once on "dataset" dict, can we train KAN on small batches of data like normally do on ANN?