Update:
EXPERIMENTAL Flux.1 Dev version added.
Please post your image, it help me understand how to train LoRA for FLUX because I can't run it before publish.
LoRA training cost Civitai Buzz. If you like my work, please consider tipping me so I can keep creating them. Reacting to my images, posting your images and using Civitai generator with my models can also help.
If you upload this LoRA to other sites please add link to this page, but please don't upload to Tensor.Art and PixAI.
You can try my models online on Tensor.Art or PixAI.
Description
This version is experimental. Trained based on Civitai guide to Flux.1
FAQ
Comments (9)
The flux Lora is amazing. Can you help me? I am trying to train Goku using civitai and the Lora is just so bad, if possible may you share 2-3 samples of your dataset's images and caption?
for flux i use 30 images with clip captions. the sample images here using the same captions in the dataset
@andinmaro146 you only use "izuku_midoriya" in the captions without any other thing?
@krigeta yes, i did not use clip captions before so maybe it is wrong
Gosh, the cost for making lora and generating pictures are really expensive O_o
Could you tell me more about flux? How it is compared to pony and how much an improvement if you think it is?
it is good at generating text and realistic images like SD3.
i don't know if it is good with anime or nsfw like pony.
it is also very expensive and need powerful hardware so i can't test if it is better but the model gallery look promising
oh my gosh! the new generation is here... just when i am enjoying pony.
but Flux just requires so much disk space. T_T
Yeah Im already at 150gb worth of stuff. Its overwhelming, pony is good enough dont feel pressured to switch
Can you train on vigilante deku? Season 6/7 deku as well