Another experimental LoRa I had flying around.Since my dataset was created using screencaps from a pretty low resolution source (what a waste of time), it may struggle with details or sharpness.Recreating the original outfit(s) worked pretty well in my opinion. I noticed some issues with double eybrows though.Just like most of my LoRas, I've trained 25 epochs, that can be requested if anyone is interested.Weight 1 is fine. Don't forget to describe the character using the appropriate tags in addition to the general activation tag.

Yui Takahashi - Tamashii: Insert

Epoch #10 -&gt; pretty clear images with high similarity to the original character. Later epochs will get more blurry and lose a lot of details.

Files with SHA256: 02b52646000ab0ffbc9324d01f9fdd4bd31f2bf54f88e64497b3b1db28d596f2