4 vector TI embedding trained on 852 images (2280 augmented) of Akai Riot's work. Trained for 67,003 steps on animefull-latest.ckpt. Higher CFG tends to give stronger lines, also can adjust strength e.g. (akai-4v:0.8) if the lines are too strong.Samples generated with Anything 4.5.I've also uploaded a <a target="_blank" rel="ugc" href="https://civarchive.com/models/9001/akai-riot-style-lora">LoRA version</a> which is a slightly stronger version of the style. I recommend having both for situations where the TI's style may be too weak or different from the original style.

Files with SHA256: 7872502e6cac1dc1b4aa927745728d4fd44ac2a005265f3b7fb18b5af4048988