Jennifer Connelly is an American actress, known for many roles in films such as Labyrinth, Career Opportunities, Hulk and Requiem for a Dream, among many others. She won an Oscar for Best Supporting Actress for her role in A Beautiful Mind (2001).
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.
Updated! Please check the "About this version" section for more info.
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 beautiful (jg_jc0n:0.99), a woman, (sfw tight fashion shirt:1.2), (long jeans), in a departmentn store, natural light, (masterpiece:1.2) (photorealistic:1.2) (best quality) (detailed skin:1.2) (intricate details) (8k) (HDR) (cinematic lighting) (sharp focus), (looking at the camera:1.1), (closeup portrait:1.1), (painted lips)"
Description
Trained on @Scottymac's colab (link here).
Number of images in the dataset: 14
Batch size: 7
Gradient: 2
Steps: 350
No captions
FAQ
Comments (13)
Jernau Gurgeh Read This
Version 2.1 is the best of the lot.
You are doing a great job JG.
Thank you
Cheers
MAX
Yep, that pretty much sums up why I wanted to create a TI of Jennifer Connelly. I wanted to recreate her looks in that specific era (late 80s, early 90s). If I want to check how she looks nowadays, more or less, every SD based model can provide me with a decent picture just by typing "Jennifer Connelly" in the prompt. It's the same with every major celebrity who was already popular in the early 2010s. Some of them are better depicted than others, though (I hate how base SD portrays Natalie Portman or Penélope Cruz, for instance).
With regards to the model, I gotta say I never really liked RealisticVision 1.3... but yeah, it might be due to a number of reasons, chief among them how each of us builds their prompts, I'd say. That has a huge effect in the final results, of course. The way I usually build my prompts, I'd say right now my favorite models are ICBINP, Avalon, Consistent Factor and Reliberate. But I'm willing to give RealisticVision another chance! If I change my mind about it, I'll let you know.
Ah, and I kind of agree that version 2.1 was probably the best, but I think it worked better with older models, whereas the new one seems to adapt better to the new models. Still, Jennifer is one of my favorite actresses of all time and, as such, I intend to achieve the best possible results with her TI. So this is still a work in progress. :)
@JernauGurgeh
I like your style and your way of thinking.
@maxhitman Wow! Man, you should write an article and publish it on this site. I gotta say I think my prompt-building idea has always been very similar to what you've described here. Still, I'll probably try with one of your examples verbatim to see how it fares with my TIs and my usual models. :)
May I suggest a slender version of her? I think she looked most beautiful in the Hulk movie.
and why no one post review? come on guys!
Might be a plan for the future. I'm still planning to improve on this version first, though!
" May I suggest a slender version of her ? "
Do you know how to write a proper Prompt script for using A.I. in Stable Diffusion ?
The correct word to use in your prompt script is....
Perfect Slim Body,
@maxhitman I doubt it will work, so I tried, and it doesn't, now feel dumb :(
@PatinaShore
I will give you one of my prompts for you to test on creating Photo-Realistic Photographs using A.I.
This special prompt was specially made to work only with Photo Realistic Checkpoints
such as Realistic Vision Versions 1.3 , 2, 3 and 4.
It may also work with other checkpoints, only IF they are for such types of realisticpictures .
Not to be used with Checkpoints such as ChillOutMix or Anime Style art and pictures.
This prompt includes this models Textual-Inversion trigger word,
it is for a Safe-for-work picture.
A Realistic Full Portrait Photo of a (JG-JC0N) woman with long dark Hair Style, Beautiful, detailed face, detailed eyes, (highly detailed skin:1.1), smile, perfect Slim body, Modelshoot style, wearing a elegant red dress designed by Versace, best quality, Professional Photography, Photorealistic, soft lighting, Sunny, standing in a forest field with wild flowers background, blurred background, volumetric fog, RAW, analog style, 8k, HD, DSLR, high quality, Fujifilm XT3, film grain, award winning, masterpiece
for your negative prompt...
((NSFW)), (Nude), (Naked), (pussy), (vagina), canvas frame,((ugly eyes, deformed iris, deformed pupils, deformed lips and teeth:1.1)), un-detailed skin, semi-realistic, CGI, 3d, render, sketch, cartoon, drawing, anime, text, monochromatic, cropped, out of frame body, worst quality, low quality, jpeg artifacts, ugly, duplicate, morbid, mutilated, extra fingers, mutated hands, poorly drawn hands, poorly drawn face, mutation, deformed, blurry, dehydrated, bad anatomy, bad proportions, extra limbs, cloned face, disfigured, gross proportions, malformed limbs, missing arms, missing legs, extra arms, extra legs, fused fingers, too many fingers, long neck, doll, child, un-detailed skin, monochrome, black and white,
Agree. She was stunning in "A Beautiful Mind" as well.
Pickle! Is this safe. Please clean this.
Safe.
Use a docker.