My first attempt at lora training, and it kinda sucks to be honest. It's difficult to get what you want and I'm not really sure what I did wrong. However, you can get some good results if you stick with it. <h1 id="parameters-gwu9xnv7b">Parameters</h1>If you do decide to use it, try 9:16 aspect ratio, i.e. 768x1344. As for the prompt, you will have to experiment, but here's what I found out:A detailed illustration = more details, but colours are a bit mutedAn illustration = less details, a little bit more colourDigital art = Same details as "an illustration" but more vibrant.Words like "Bold" or "Vibrant" will add some more colour into it.<h1 id="will-there-be-updates-qygrs86sc">Will there be updates?</h1>I'll try to fix it, but the updates will depend on my luck. This is really hard and expensive/time consuming.

Files with SHA256: 941436bf851f4c747c5159c723dba6a1596df13897efbddaa5dcce814b815408