hi! i thought i was gonna win the bounty for this and checked ur version 😭😭 can u tell me how u did it?? i used hollowstrawberry's lora trainer XL, trained it on illustrious, 1 activation tag, 2 repeats, 20 epochs, saves 10 last epochs, learning rate 5e-5, 0 for text encoder, cosine with restarts, scehduler number 3, 0.05 warmup ratio, min snr gamma enabled, value 8, ip noise gamma off, multinoise on, lora structure, 16 net dim, 8 net alpha, 16 conv dim, 8 conv alpha, and 1 train batch size