Depending on the base model,her eyes will vary from either small,large,wrong color or her pupils will just be gone.The lora doesn't handle the looking at viewer prompt well but does work the facing viewer prompt.That also depends on your base model choice.I use only one base model and I'm determine to get it to work.
Not sure if its because of anime but using these for the negative prompts helps greatly: EasyNegativeV2, worst quality, low quality, watermark, signature, simple background, bad anatomy, bad hands, deformed limbs, blurry, cropped, cross-eyed, extra arms, speech bubble, extra legs, extra limbs, bad proportions, poorly drawn hands, text, flat background,modern, recent, old, oldest, anime, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, bad hands, missing fingers, extra digits, fewer digits, cropped, very displeasing, , bad anatomy, sketch, jpeg artifacts, signature, watermark, username, simple background, conjoined, bad ai-generated
I have alot of loras so adding this info wasn't easy.
