Noticed that most models produce either non-Indian looking women, or if prompted, produce very dark skin tones, which some, but not all Indian women have. This textual inversion tries to address that. This model produces photo-realistic women with Indian skin-tone, facial features, and body types.

Produces NSFW images unless used with very low weightages, or forced through prompts like ((clothed)).

Will likely follow up with a SFW/ mixed version.

his is a Textual Inversion I have created, this is mainly useful to improve your overall aesthetics and character anatomy.

I don't really have a suggested weight range for this, because it really depends on what model, lora and how complicated your prompt will be. These elements can effect how much weight you will need, I usually start without any weight just to see how much it effects my generations.

The images I have generated are purely just prompts and my Textual Invertion, I only switch around the checkpoint model to test the effect.

You can try more different variation and you can get even better results then I do, as this was just some testing.

A commission of @Zorglub.

Shania Twain is a Canadian singer and actress known as the Queen of Country Pop. She's one of the best selling musicians of all time and a winner of 5 Grammy Awards.

1500-step TI trained on a dataset of 18 images with my usual settings.

Appreciate my work? My TIs are free, but you can always buy me a coffee. :)

Curious about my work process? I have summarized it here.

How to create a good prompt using my TIs

You're obviously free to experiment, but bear in mind that my TIs are trained with a more or less fixed phrasing, that normally starts with:

"photo of EMBEDDING_NAME, a woman"

So I recommend always starting your prompt like that and then building the rest of the prompt from there. For instance, "photo of (shtw41n:0.99), a woman, RAW, close portrait photo, leopard print suit with hood, leopard print top, pale skin, slim body, (high detailed skin:1.2), 8k uhd, dslr, soft lighting, high quality, film grain, Fujifilm XT3 sharp focus, f 5.6, redhead, (hood on:1.2)".

Avori is a full-time Video Game Streamer and Content Creator with over 1.7M followers across Twitch, Facebook, TikTok and Instagram.

I found her through a gif that was posted in a discord conversation and decided to create a ti.

Still learning and improving the progress of the training, any feedback is welcome.

Feel free to leave me a tip if you like what I am doing :)

Embedding of an original character. This person does not exist.

A request of @lemke.

Judith Rakers is a German journalist and TV anchorwoman, mainly known as the host of the news programme Tagesschau on ARD.

1500-step TI trained on a dataset of 18 images with my usual settings.

Appreciate my work? My TIs are free, but you can always buy me a coffee. :)

Curious about my work process? I have summarized it here.

How to create a good prompt using my TIs

You're obviously free to experiment, but bear in mind that my TIs are trained with a more or less fixed phrasing, that normally starts with:

"photo of EMBEDDING_NAME, a woman"

So I recommend always starting your prompt like that and then building the rest of the prompt from there. For instance, "photo of (judrak:0.99), a woman, RAW, close portrait photo, black leather jacket, tank top with logo, pale skin, slim body, (high detailed skin:1.2), 8k uhd, dslr, soft lighting, high quality, film grain, Fujifilm XT3 sharp focus, f 5.6, long blonde hair".

This ti was somewhat of a challenge, a friend of mine told me I couldn't do a ti of a female bodybuilder look good. I disagree and here are the results.

Lean Beef Patty, leanbeefpatty on instagram, is an instagram model and fitness advocate with over 4.6 million followers.

This was quite the fun challenge (not challenging tho) to do and I hope that there a a few people on here that find enjoyment with this TI.

Still learning and improving the progress of the training, any feedback is welcome.

Feel free to leave me a tip if you like what I am doing :)

I trained this model with 10.000 steps (v1.0) and 20.000 steps (v1.1), I'm no expert, but I hope, you like it.
Add into your prompt: "(ElviraMistress:0.75)"
Trained with 512x512 pixels.

I accidentially deleted version 1.0, re-uploaded, sorry :)

Emily O'Hara Ratajkowski (born June 7, 1991) is an American model and former actress. Born in London to American parents and raised in Encinitas, California, Ratajkowski began her career as a child actor appearing in two episodes of the Nickelodeon series iCarly (2009–2010). Her modeling debut was on the cover of the March 2012 issue of the erotic magazine treats!, which led to her appearing in two music videos: Robin Thicke's "Blurred Lines" and Maroon 5's "Love Somebody".

If you like my work and you feel like it, you can invite me to a Ko-fi!

Created this as my first ever TI as a test as I plan on making more in the future and it turned out well in the end so figured I would put it here for people to use. Mind you this is a very alpha build for the first one and will be posting a newer version at some point.

The Model I used is one I merged myself using Av3, Av3 pruned, and another I can't remember. I might post it on this site at some point, just not sure if I can.

prompt used:

mecha, mechanical arms, mechanical halo, mechanical wings, mechanical parts, intricate mechanical bodysuit, mecha girl, hair ornament, earphone, (mecha wings:1.4), (cybernetic:1.10), (intricate mecha body:1.07), (headgear:1.2), neon trim, glowing neon lights, red, mechanical halo, mecha knee pads, wristband, (mechanical shield:1.09), axe, (mechanical axe:1.09), corruption tatoo, cyborg, robot joints, cybernetic, science fiction, cyberpunk, machinery, mechanization, outer space, black hole, 1girl, (multicolored hair, braid:1.2),

Embedding of actress Sophia Bush, known for her cheekiness and her raspy voice. You might have seen her in series such as One Three Hill or Chicago PD.

512x512 or 512x768 works best.

30ish steps DPM++ 2M KARRAS

CFG 7-8

I recommend not using hires fix but using the new Controlnet 1.1 / Ultimade SD upscaling with tiles, as it preserves her dimples and face complexity better. Hires fix tends to make her face look too even which takes away from the likeness.

Riley Guava was on Instagram, Onlyfans, TikTok, but she didn't exist.

She was a fake (AI-generated/photoshopped) model.

Now can YOU play with her.

A commission of @psytrancehero.

Anna Tatu is a Czech adult model who has featured in several movies and magazines produced by Playboy.

This is a 1200-step TI trained on a dataset of 12 images with my usual settings.

Why 12? Because I tried first with my usual 18 and I couldn't get the eyes right. I trained up to 4200 steps (probably my record high in the last couple of months) and... well, there was no fixing it, so I went back to the dataset, removed 6 suspicious pics and didn't add any new ones. Then retrained from scratch and I'm pretty satisfied with the final result, although I may keep working on it at a later point. Oh, and fair warning: this TI is very NSFW, given the source material, so if you don't want to see any nudity, please prompt accordingly.

Appreciate my work? My TIs are free, but you can always buy me a coffee. :)

Curious about my work process? I have summarized it here.

How to create a good prompt using my TIs

You're obviously free to experiment, but bear in mind that my TIs are trained with a more or less fixed phrasing, that normally starts with:

"photo of EMBEDDING_NAME, a woman"

So I recommend always starting your prompt like that and then building the rest of the prompt from there. For instance, "photo of (4nn4t4tu:0.99), a woman, RAW, close portrait photo, black leather jacket, tank top with logo, pale skin, slim body, (high detailed skin:1.2), 8k uhd, dslr, soft lighting, high quality, film grain, Fujifilm XT3 sharp focus, f 5.6, long blonde hair".

This textual inversion was trained off images of the western wall of Room 57 of the British Museum.

Levant_57 is the text prompt to call.

I prefer DDIM Sampler with CFG Scale 8

Negative prompts: "people, group, crowd, man, woman," help to make (but won't ensure) you not get any people generating in the images.

This is a draft version of a textual inversion for Erin Gray aka Colonel Wilma Deering from the 1980s Buck Rogers TV series.

It generates a fair likeness but it's a bit undersaturated due to poor image sourcing. I think I can improve with better images and/or processing the images a bit, so this is released as version v0.5.

Tina Kunakey

Tina is a French model and actress.

This is a textual inversion, if you'd rather have a LoRA, here you go!

You can learn more about my process in my blog post HERE

I can create a custom Textual Inversion or LoRA just for you!

Do you like my work? Buy me a beer!

[ SDv1.4, 4e-03, 4700 steps, v/s .200 ]

Yoon Young Bae

Bae is a South Korean fashion model.

This is a textual inversion, if you'd rather have a LoRA, click me.

You can learn more about my process in my blog post HERE

I can create a custom Textual Inversion or LoRA just for you!

Do you like my work? Buy me a beer!

[ SDv1.4, 4e-03, 4700 steps, v/s .200 ]

Sonia Ben Ammar

Sonia is a Tunisian model, singer, and actress.

This is a textual inversion, if you'd rather have a LoRA, its right here.

You can learn more about my process in my blog post HERE

I can create a custom Textual Inversion or LoRA just for you!

Do you like my work? Buy me a beer!

[ SDv1.4, 4e-03, 4200 steps, v/s .200 ]

Monica Alvarez

Monica is an American model on the Instagram and the TikTok.

This is a textual inversion, If you'd rather have a LoRA, check it out here.

You can learn more about my process in my blog post HERE

I can create a custom Textual Inversion or LoRA just for you!

Do you like my work? Buy me a beer!

[ SDv1.4, 4e-03, 4200 steps, v/s .201 ]

This ti was requested by my fellow ti creator, friend and person I enjoy talking to, JernauGurgeh.

Tammy, or tammyhembrow on instagram, is an instagram model and fitness advocate with over 16 million followers.

First time doing a ti soley on instagram pictures. I hope you enjoy this ti as much as I do!

Still learning and improving the progress of the training, any feedback is welcome.

Feel free to leave me a tip if you like what I am doing :)