It is not perfect but it is a decent v1. This model was trained on 331 images of naked women with various body types and ethnicities. It is a combination of my datasets for breast sizes and innie pussies.
I trained with the below captions as triggers but it works well without them.
Triggers: She has [b cup | c cup | d cup | large perky breasts | a flat chest]. She has an innie pussy.
LTX-2 Model of Mickey Mouse: https://civarchive.com/models/2325307/the-mouse?modelVersionId=2615823
Description
FAQ
Comments (9)
Thank you, King.
LTX-2 is just garbage, don't waste your time on this nonsense. Just forget it.
Yikes, how the hell is this the "improved" anatomy? LTX is way undercooked to be in the public right now
It reminds me of flux... Equally poor nudity, no matter how many datasets are used for loras, cannot represent the appearance of nudity. wan is much better
pretty freaking amazing.
Fantastic first go at it! - no doubt v2 will be jUST right! If anyone can do it it's you PLAYTIME. Btw, this model is a game changer, the movement is so refreshing, try adding more fast dramatic camera direction, the flow is great
Does it works with gguf (4q k m)model?
Is this for T2V aswell?
I am curious how you trained vulva. I am trying to train a lora and vulvas are not working as well. Did you use the heretic text encoder, or did you caption it a certain way?
Any tips would be helpful. Thanks.
Details
Files
LTX-2 - Better Female Nudity.safetensors
Mirrors
LTX-2 - Better Female Nudity.safetensors
LTX-2 - Better Female Nudity.safetensors
102-Better-Female-Nudity-jan27.safetensors
LTX-2 - Better Female Nudity.safetensors
LTX-2 - Better Female Nudity.safetensors
LTX-2 - Better Female Nudity.safetensors
LTX-2 - Better Female Nudity.safetensors
LTX-2_-_Better_Female_Nudity.safetensors