Welcome to my League of Legends Splashart Style Model
Let me start this off by saying that this is my very first Lora, so any feedback (positive AND negative) is much appreciated.
I would also love to see what you all are creating, so feel free to post whatever you generate
Trained on 144 League of Legends Splash Arts
__________________________________________________________________________________________________
General Triggerwords
League of Legends Splash Art
Dynamic
Specified Triggerwords
Arcana, Battle_Academia, Blood_Moon, Coven, Dawnbringer, Empyrean, Firecracker, Fright_Night, KDA, KDA_ALL_OUT, Lunar_Beast, Mythmaker, Porcelain, Psyops, Shan_Hai, Snowmoon, Soul_Fighter, Spirit_Blossom (Akana / Kanmei), Winterblessed
Yordle (Just generates cute animals, not always specifically Yordles)
If you are using any of the Skin Line specific triggerwords make sure to describe the Skinline itself, these dont fully work on their own 100% of the time sadly.
__________________________________________________________________________________________________
Usage Tips
Use this Lora at a Weight of 0.3-0.7 for the best Results, you can play around with it a bit but at a Weight of 1 it tends to look quite overbaked.
I would HIGHLY RECOMMEND using the suggested Resources to use this Lora. This Lora sadly struggles a bit with hands and that specific combination of Negative Embeddings and the Better Hands Lora helps eliviate that by quite a lot.
Keep in mind that the creator of the EnvyBetterHands LoCon recommends against any Bad Hands Embeddings but recommends using "nice hands, perfect hands" in the Positive Prompts and "(extra fingers, deformed hands, polydactyl:1.5)" in the Negative Prompts.
I would also strongly recommend considering AfterDetailer for the faces.
I would also recommend using RevAnimated or similar Models for consistent results.
Other Models CAN work but the results may vary, i have found that Realistic Models don't do that well.
This Lora works well using 512x512, 768x512 and 512x768.
I would highly recommend using 2x Highres. fix using the Latent Upscaler at a Denoising Strength of 0.5-0.8 since this Lora struggles with details at lower resolutions.
__________________________________________________________________________________________________
Feel free to leave any questions you might have in the comments below and I'll get to you as soon as i see them.
Description
FAQ
Comments (4)
I've updated the prompting style, it is more simple now but results in much better hands, all the negative embeddings and loras are listed under Suggested Resources <3
Hello, I would like to thank you for this lora. It's one of the best lora in terms of splash art I could get since one year.
I can generate gorgeous images with my combination merge and this lora.
I hope you gona improve in future in terms of quality and colors :)
Dont give up
is there any place i could get some help to get this to work with similar results as you show in the examples? im new and there is SO much information, i dont know what parts im missing and what to focuz on. my results dont look like yours
Sorry about the late response
About getting similar results to mine, im using a decent amount of embeddings (They are all under suggested resources), and a Lora to fix the hands since this Lora really struggles with those
If you do use all of those together and it still does not look similar it might either just be the Checkpoint Model itself or possibly the prompting Style
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.