100th Lora! thanks for sticking by!
Weight 0.8 - 0.1
"TRLCLAS, brown hair, brown eyes, single braid, lips, teal tanktop, large breasts, sleeveless, shoulder strap, brown shorts, thigh holster"
Handpicked dataset with a bunch of iterations of classic Lara in mind.
Her face isn't as malleable as it was in 1.0 but it looks much more accurate and produces good gens at lower weights.
Use "hip vent" and "leotard" to change the shape of her tanktop
Use "round tinted eyewear" for her classic glasses, also trained to use "looking over eyewear"
Gonna take it a bit easier after this. I'll drop most of the stuff that's already made and after that I'll probably go back to my slumber for a bit. Thanks for all the downloads and gens so far!
Description
Weight 0.8 - 0.1
Lora trained with her classic outfit in mind, using a handpicked database.
Use "red-tinted eyewear"and "sunglasses" for her usual glasses.
Use "3D" as a negative prompt.
If you are getting bleeding from her backpack straps, using "straps" and "harness" usually helps
So turns out that last Lora was a bit of a mistake and I was a bit frustrated by it. I decided to rework the database and train it with Pony instead and I can say that I'm significantly happier with the results now.
On another note, there'll be more Pony stuff rolling out eventually. However, they will release significantly slower compared to my other Lora. If there's any of my older works that you'd like to see updated like this, I'd appreciate if you drop a comment letting me know which one.
FAQ
Comments (5)
Would it be too much to ask for this style to be added to your 1.5 Lara?
What do you mean by style?
this looks more 2d than the images for Tomb Raider I Classic Lara Croft. Your images there are 2.5d-ish I think? I don't know, in truth I've yet to try it and it could be the prompts that make the difference. Regardless, great looking either way and looking forward to trying it out
@doveship I honestly think it is just kind of a thing with SD 1.5 and "realistic" characters. If you look at my B. Jenet Lora as an example, it looks a bit 3D but since the character is much more stylized and has more cartoony proportions, I feel like it is easier to train with an anime model. The thing is, the 1.5 model is trained with a ton of 3D model screenshots and it has the exact same database as this one. My guess is that it is just a result of Pony training images at a much higher resolution. I feel like the very first Lara model looks like that because the database was significantly smaller. The newer ones are trained with Animefull and the oldest one was trained with pixai with whatever model they use for anime.
Might be worth giving it another go and training it with a different model, I won't revisit it for a while though. I don't want to saturate everyone with Lara Loras. Sorry for the long rant btw.
@Lamancha50002310 No apology necessary, a very thorough response thank you
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.