<h2 id="heading-1298">V2.0</h2>Two gothic/gothic lolita text inversions<ul><li>b_horror_v2_1 is cute and dressy, good for gothic lolita and cute vampires</li><li>b_horror_v2_2 is dark, good for scary vampires</li></ul>Prompt:<ul><li>nsfw</li><li>backlight (all the training samples have a girl dressed in black with a very bright moon or a very bright window)</li><li>colorless (most drawings have low contrast or low saturation)</li></ul>Other notes:<ul><li>this was difficult to train, mostly because all source drawings have low contrast, low saturation, lots of small elements that turn into artifacts, and because most girls have a black corset, the waist area is always excessively dark/shadowed</li><li>correcting brightness/contract, maybe this: <a target="_blank" rel="ugc" href="https://civarchive.com/articles/1125/new-tool-cdcolordetail-tuner">https://civarchive.com/articles/1125/new-tool-cdcolordetail-tuner</a></li><li>therefore, a good way to make a drawing without the text inversion is to add "victorian dress, frills, backlight, low contrast, low saturation, 90's style manga" to the prompt</li><li>the model has a tendency to:<ul><li>make girls that hold things in their hands (their dress, cake, flowers, a glass of blood)</li><li>add some chibi clones (sometimes)</li></ul></li></ul><h2 id="heading-1299">V1.0-Experimental</h2>I was doing sth with the computer, and this appeared.You need to play with CFG scale, embedding strength, and strength of 'nsfw' in the negative.Embeddings:<ul><li>b_horror_1: vampire, red and black</li><li>b_horror_2: demon, pink and black, scary symmetry</li><li>b_horror_3: nsfw, lots of frills, good for goth lolita</li></ul>Other notes:<ul><li>horror_sth_3 is a mix of 20% horror1, 50% horror2, 30% horror3</li><li>I have included some images with latent upscale, you should try to obtain the same results by hand (it's how intermediate checkpoint examples are generated when training a text inversion)</li></ul>

B-Horror (~ Japanese Victorian Gothic)

b_horror_3.pt

Files with SHA256: fbdd941751cb99dff37b1f7e868e409a4a2c5d45ef71a6bd8b3e2f41ec34faa0