Demonstrating SimpleTuner's AnyFlow support by distilling Anima v1.0 using e621 dataset. Recommended weights for inference are dmd-v3 or dmd-v5.