A request of @Bzas76. Sorry I couldn't work with an older dataset, mate, but I hope you'll like this one!
India Reynolds is a British model and social media celebrity, known for her recent stint in Love Island, among other things.
Update: Retrained from scratch, since my work process has changed quite a bit since I did this TI. I think the previous version, even if it was only a 200-step one, was very good, but perhaps slightly less accurate than desired. India looked gorgeous in almost every pic, though, which is why I was hesitant to update the model. I think this new version (2.0) captures her looks better without losing terrain in terms of beauty.
This new version is a 1000 step TI, trained on a dataset of 18 images.
Appreciate my work? My TIs are free, but you can always buy me a coffee. :)
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 (1nd1areyn0lds:0.99), a woman, RAW, close portrait photo, long brown coat, turtleneck sweater, 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".
Description
FAQ
Comments (2)
I'm in awe of your embeddings, JernauGurgeh! Mine never come out even close to this good.
Any chance you could share your training settings as a couple of screenshots from whatever repo you're using to train these? I realize that settings need to change a little depending on what dataset one has, but some template to start the experiments with would be a huge help.
Also, do you have a digital tip jar somewhere that I could put a coin in? I feel since I get so much benefit from your embeddings, the least I could do is buy you a beer or two.
What kind of witchcraft are you using with only 200 steps?!
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.