Danielle - New Jeans

Danielle is a South Korean-Australian singer, a member of the girl group NewJeans.

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, 4500, v/s .200 ]

Katheryn Winnick (born December 17, 1977 is a Canadian actress. She is known for her starring roles in the television series Vikings (2013–2020), Wu Assassins (2019), and Big Sky (2020–present), and her recurring role on the television series Bones (2010–2011). She also starred in the films Amusement (2008), Choose (2010), A Glimpse Inside the Mind of Charles Swan III (2012), The Art of the Steal (2013), Polar (2019), and The Marksman (2021).

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

Key word: alissa white-gluz

Using realisticVisionV20 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Jewel Staite is an actress from Canada. You might have seen her in the series "Firefly" or "Star Gate Atlantis".

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 :)

Just an attempt to train the 80s model "Maria Whittaker" as textual inversion.
Have fun!

Stefania Liberakakis is a Greek-Dutch singer, voice actress and YouTuber.

Used model: CyberRealistic V2.1

Katy M o'Brian. She's been in The Mandolorian, Ant-Man and the Wasp: Quantumania, The Rookie and some other stuff.

I just thought she would make a great character for Cyberpunk themed images.

Trained on 29 image, 16 vector per token, used BLIP for captions (without checking results/lazy), 11,750 steps.

Don't know if it makes a difference but it was set to v1-5 when training but I use cyberrealistic_v13 for generation.

If you know better settings I should be using, please let me know.

Instagram model @amandaeliselee

[This is a request] Cristina Scabbia from the band Lacuna Coil

Key word: Cristina Scabbia

Using Chilloutmix for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

This is a textual inversion model of fictional character "Sara". Just place in your "embeddings" folder and use word "saraVoxelsd15" in your prompt.

Key word: Supernatural jack

Recommend to use "Restore face" if you get a bad result.

Using RealisticVision v2 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Key word: Supernatural charlie

Using Deliberate v2 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Key word: Supernatural bobby

Recommend to use "(Supernatural bobby:0.6)" higher will get a bad result may be diffrent depending on checkpoint used. Poorly trained model due to poor quality of training material. But still wanted to make a model of him.

Using RealisticVision v2 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

19 year old Swedish nude model.

A commission of @Zorglub.

Ming-Na Wen is an American actress of Chinese descent, known for her roles in many movies and different TV shows, including The Mandalorian and The Book of Boba Fett.

This TI trained really fast, and I ended up selecting step 600 as the best one out of the handful I tested. As usual, it was trained on a dataset of 18 images and with my standard 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 (m1ngnawen-600:0.99), a woman, RAW, close portrait photo, bodysuit with zipper, cleavage, fishnet stockings, 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"

follow me on Instagram, Youtube, Patreon and my website.

Meg Ryan

Suggested Weight: 1.0

Suggested prompts

• professional portrait photo of 20yo m3gry4n as an astronaut, asymmetrical bob haircut, light bokeh, elle magazine

• adding "young" or an age like "20yo" / "24yo" helps get a more youthful look

Training

• trained on ~16 images

• Learning Rate: 0.005:1000, 0.001:2000, 0.0001:5000, 0.00005

• Total Steps: 15,000

• Batch Size: 1

• Gradient Steps: 1

RoxieCipher is a person who doesn't exist. You can use her if you need a consistent character.

Please have a look at other Nobody models for contributions by other folks including those created with the LastName and NBDY tags.

Follow me if you'd like to be notified of future models!

Corina Caragea is a Romanian news anchor and can also be found on Instagram.

She has over 476 thousand followers.

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 :)

This is an experiment to see if I can make a TI embedding that gets the flavour of konyconi’s BohoAI LORA.

https://civitai.com/models/51966/bohoai

Thank you to @konyconi for sharing his dataset for the excellent BohoAI LORA.

https://civitai.com/models/52697/tutorial-konyconi-style-lora

The showcase uses 2 models:

revAnimated_v122.safetensors [4199bcdd14] with clip skip = 2

avalonTruvision_v2.safetensors [a4df55d292] with clip skip = 1

This TI can produce some decent Boho pix but it gets confused sometimes... eg asking for a spaceship and getting a truck. Perhaps for this sort of TI you need to use a lot more pictures in the training dataset, with more subject variation?

---------------------------

Update 09 May 2023

Continued the training to step 4000, and then 5000.

kcboho07-4000 produces a stronger Boho style.

kcboho07-5000 is stronger again but has increased duplication/repetition. e.g. more fingers, more hands, duplicate cities floating in the sky.

Tried 6000 steps but it’s even worse - overcooked.

I’ve uploaded the 4000 step version as, probably, the best result for this experiment.

Also uploaded the 5000 step version since it can produce nice results with careful object prompts.

---------------------------

I’ve been struggling to work out how to make a style TI...

what makes a good training dataset?

what training settings should I use in automatic1111?

how long to cook the TI for?

For my training dataset I copied konyconi’s 76 1024x1024 images to a new folder without the associated TXT files, and reduced them all to 512x512. Then I renamed them “01 aeroplane.png”, “02 city.png”, “03 tank.png” etc.

Why? Because I was trying to match what I’ve done in the past for TIs that ended up usable. The reduced images dataset folder is what I used in the settings below.

The wikipage for automatic1111 Textual Inversion is here:

https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Textual-Inversion

but way out of date. Last rev Jan 5, I’m writing this May 8.

I found this thread useful in parts. It’s a long read!

https://github.com/AUTOMATIC1111/stable-diffusion-webui/discussions/1528

Training model: v1-5-pruned.ckpt [e1441589a6]

I used this because I don’t know any better, and it’s been useful in the past. Should I be using a different model for training, or is the base SD15 the best thing to use? No idea.

Create embedding:

name: kcboho07

initialization text: boho style photo

number of vectors per token: 4

Train embedding:

Embedding name: kcboho07

Embedding Learning Rate: 0.001:250, 0.0005:500, 0.00075:1000, 0.001

Gradient Clipping: disabled

Batch size: 1

Dataset directory: wherever you’ve put it on your computer

Log directory: textual_inversion

Prompt template: minimum_style_2.txt

The template has 3 lines:

<<<

[name] style, [filewords]

[name] style, a photo of [filewords]

[name] style, an illustration of [filewords]

>>>

Width = Height = 512

Do not resize images: OFF

Max steps: 3000

Save image steps: 25

Save embedding steps: 25

Use PNG alpha channel: OFF

Save images with embedding in PNG chunks: ON

Read parameters from txt2img tab: OFF

Shuffle tags: OFF

Drop out tags: 0

Latent sampling method: deterministic

Training time: about 50mins per 1000 steps on a 2060/6GB.

The TI at 3000 steps does produce a Boho style, although I think it’s a bit hit-and-miss compared to the BohoAI LORA.

If anyone has suggestions about what I should be doing differently please add a comment. Or if I’m doing anything obviously stupid! :-)

A commission of @findingavarice521.

Madison Ivy is a German/American pornstar and adult model.

1000-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 (m4dison1vy:0.99), a woman, RAW, close portrait photo, bodysuit with zipper, cleavage, fishnet stockings, 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"

Android 18 kisses Tenshinhan

follow me on Instagram, Youtube, Patreon and my website.

Hedy Lamarr

Suggested Weight: 1.0

Suggested prompts

• 1988 color horror movie poster of h3dy14m4rr

• add "color" to the prompt (color photo, color poster, etc) otherwise it will likely be b&w

• The hair style is somewhat baked in, to get straight hair you will need to strongly emphasize that in the prompt like ((straight hair))

Training

• trained on ~23 images

• Learning Rate: 0.005:1000, 0.001:2000, 0.0001:5000, 0.00005

• Total Steps: 15,000

• Batch Size: 1

• Gradient Steps: 1

Zelina Vega / Thea Trinidad

Thea is an American professional wrestler and cosplayer also known in the WWE as Zelina Vega. This embedding is very finicky about models, I found dreamlikephotoreal to create the best likeness.

Also, this is a work in progress. It's getting better and I've been asked to upload this version, so there 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, 5200, v/s .199 ]

Key word: Supernatural castiel

Recommend to use "Restore face" if you get a bad result.

Using Deliberate v2 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Key word: Supernatural lucifer

Using Deliberate v2 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Key word: Supernatural crowley

Recommend to use "Restore face" if you get a bad result.

Using Deliberate v2 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Danielle Kristie Sharp (born 21 August 1991) is an English model, actress and video games streamer.

Requested by Bzas.

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

A very versatile negative embedding that helps generate high quality images. Higher chances of producing normal looking hands in realistic images.

Best Practices:

Keep your token usage to below 75 for best effect. Most of the time, you don't need to add other quality specifiers to your prompts (e.g. worst quality, bad quality).

Usage:

Negative prompt: bad_pictures

If you notice any artifacts, you can try increasing or decreasing the weighting like this: (bad_pictures:1.1) or [bad_pictures:0.8].

Values between 0.8 and 1.2 seems to work best (clip size 2):

Tested on:

• VersaMix (https://civitai.com/models/20865/versamix)

• Lyriel (https://civitai.com/models/22922/lyriel)

• Anything V5 (https://civitai.com/models/9409?modelVersionId=30163)

this is my first man proje, hope you guys enjoy.

要不要凑个CP弄个井上真央？

27 year old Czech/American TikTok/Instagram model and business woman.

Key word: Supernatural death, (male, age 60:1.4),

Using realisticVisionV20 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Key word: Rowena

Using realisticVisionV20 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Key word: Supernatural sam

Using realisticVisionV20 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Key word: Supernatural dean

Using realisticVisionV20 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Key word: Tarja Turunen

Using realisticVisionV20 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Key word: Gwen stefani

Using realisticVisionV20 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

Key word: mix goths

Using realisticVisionV20 for all the photos.

DMP++ 2M Karras

20-30 steps

Use as you wish. Asking permission for AI generated material is stupid. Since everything is taken from someone else. Long live pirates! F them greedy..

A commission of @psytrancehero.

Katya Clover is a Russian pornstar, adult model and social media personality.

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 (katclov:0.99), a woman, RAW, close portrait photo, bodysuit with zipper, cleavage, fishnet stockings, 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".

Laia Manzanares Tomàs is a Spanish actress best known for her role as the cheerleader and love interest in Tame Impala's 2015 music video for The Less I Know the Better, which became a sleeper hit and cult classic in the late 2010s. She is also known for playing the role of Ukrainian student Oksana Casanoves in the Spanish television series Merlí.

Requested by JernauGuregh.

You can invite me to a Ko-Fi if you want!