(nyahility956) (2024)
Image description: A woman with horns lies on a traditional Japanese tatami mat, surrounded by scattered cherry blossom petals. She is dressed in a dark, elegant kimono with intricate designs, and the garment is loosened to reveal her shoulders. The light highlights the textures of the tatami mat and the person’s attire.
Full Generation Parameters:
anime screencap, Japanese woman lying on floor on her back, black lips, black kimono with white obi, ((white Oni horns)), black hair, white cherry blossom petal hair accessories, 8k, intricate detail, hands near head, white tatami floor, looking away, cinematic lighting, amazing quality, amazing shading, black eyes, extremely detailed eyes and face, ((open shoulders, collarbone)), beautiful, detailed eyes, ((masterpiece, best quality)), ((full body, far shot)).
Negative prompt: (low quality, worst quality:1.4), negativeXL_D, cgi, text, signature, watermark, extra limbs
Steps: 30, Size: 1024x1024, Seed: 157178864, Sampler: Euler a, CFG scale: 7, Clip skip: 2
Oh hey, I just recognized your username from the other day with the user that kept deleting his messages, ha.
Anyways, what I mean is the model name like sdxl or juggernaut, etc. It should be in the metadata.
I’m not complaining though, just getting the prompt alone is a blessing.
It’s hidden in a spoiler right under the text description.
Yea, it’s just missing the checkpoint used. It’s under resources on civitAI. I thought you were grabbing the image metadata where it shows up as “Model:SD_XL” for example. But like I said, it’s not really a problem, especially when I just realized you link every image as well which ultimately solves it.
I didn’t realize. It has AAM XL (Anime Mix) as a linked resource, I hadn’t realized the parameters didn’t have the model name in it.
Yea it’s a bit of an odd choice on civitAIs part. They dump Loras and embeddings in there too so it can become a bit messy.