Training the velocity network by regressing on random straight-line slices, without ever running the full solver during training.
Continue to AI University →