A concept LoRA trained on a lot of recent splash art images from the game League of Legends
Description
This is a test version made up for anima-preview-2, neither this is the final version of this model for the preview 2, nor the final version for anima itself.
Training took both way more time to reach good results than sdxl and several weeks to actually get good training params, and yet this is not where i'd like it to be, so expect updates.
Experiment with it, check the results, and if possible give feedback.
FAQ
Comments (4)
MAKE ANIMA # 1️⃣
ILLU R.I.P. 🪦🪦🪦
Wasn't sure what this was going to do but it's actually great for making dynamic widescreen shots. While it does lean towards a specific style and face type, it seems to respond very well when stacking an artist tag or style LoRA even without reducing the strength of this LoRA. Here's an album with some quick tests.
If this version still isn't where you want it to be then I can't wait to see a version that you're fully satisfied with. Thanks for making this!
Thank you, this was trained on the preview version of anima, so not only from my side, but also from anima team side, there's a lot of improvements to be made
thanks a lot,great lora for LOL style