Recommend to use with this checkpoint
0.5 is a good starting value.
Negative Realistic is recommended for most broad use in Pony
1.0 Lora is only recommended for up scaling, and realism.
Thank you for continuing to support my training addiction with your tip (By default I know) of just 1 buzz per generation while using this lora.
An example of my cost:
11¢ per hour for training power consumption
15¢ per hour for amortization of equipment cost over 3 years
Paid time on this model would be around $350
Description
FAQ
Comments (9)
would you consider trying one for hairy legs for futas/women?
You might be able to get that through prompting, I'm still in Comfy I will give it a shot
@Felldude you can but it isn't consistent, sometimes it forces men into the picture when you do, which is why i was thinking a lora might help to stabilize the concept
I wasn't able to reliably prompt anything, it would be a difficult training as the I would assume it would tend toward masculine legs even if you had feminine legs in the training data.
It would require a large training set and possible specific block data.
I would consider it but someone who has trained futa or masculine subjects might be a better candidate for that training
@Malessar we got the same resault
Thanks for this! Is ponyXL_FallenOne your own creation, and is it uploaded anywhere?
It is a training merge I am still working on. It isn't currently available. I plan to post another example image with documentation of my progress with it later on. (Getting 2k results finally)
@Felldude thanks for the response! any plans to release it soon for wider testing?
@disapproving_daddy Possibly, without the model being sponsored it is unlikely garner much interest even if it can native generate at 1600 and upscale to 2k