Experimental, first flux try testing civitai lora trainer, trained on 19 close-up photosTrigger word: doja, you may want to specify as doja woman, not doja cat ,if you don't want to see an animalrank: 32 (probably overkilled )detailed tag, does simpler tag perform better? , hmmtraining resolution: 512, but inference at larger res is fine as shown in examplenot being able to train on multi-res is a regret on civitai trainer, hmmcannot test training regularization images properly here as well, hmm

Doja Cat - Flux.1 [dev]

rank: 32 (probably overkilled )engine: kohyadetailed tag, can simpler tag perform better? , hmmtraining resolution: 512, not being able to train on multi-res is a regret on civitai trainer

Files with SHA256: 7809d4eac3d58eafb5163d5bf2a68e2c621105805f0127cae7509ecfbc4c21e6