Genshin TCG Style [LTX-2.3]
https://civitai.com/models/1768496/genshin-tcg-style
Trigger Word: Genshin_TCG
Model: anima-base-v1.0
Diffusion-pipe was used for training
The workflow is in the images themselves
Training details
The same dataset used for the video was used for training. However, I decided to add every frame from the video to the dataset. I'm not sure this was a good idea, but it definitely provided good data augmentation.
The best lora was taken from Epoch 9, which was version 13 with a rank of 32 and lr=5e-6.
Visualization of all learning options. Perhaps this information will be useful to someone.





















![Genshin TCG Style [Anima]](https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/cd55bfc1-3880-40ad-819b-252620382bb9/width=1600/v_00001_.jpeg)