A nude/NSFW capable model that focuses on images with a female primary subject.
It is heavily biased towards nude images.
Recommended Settings
Sampler is dpmpp_2m (DPM++ 2M) and the scheduler is beta
No VAE or CLIP is baked in. Use separate sources for those.
Each sample/preview image contains the used workflow. Here's a quick article with simpler more beginner-friendly workflow. This is a recommended starting point.
FP8 Download
For some reason, when you upload multiple formats (BF16 + FP8) on one model, which they now support, they hide the FP8 under the BF16 model group in the side menu...
Early Access is enabled to support the development of new versions and models.
Description
FAQ
Comments (25)
wish i had the 5k buzz required to start making content with this model. i guess a 14days cool off isn't that far away.
You can mess around with 4.0 meanwhile. It's also very capable. But may lack lizards on body.
Could you please release the fp16 version of newer models?
Hmm, maybe, my core is already fp8, so I would have to merge in an fp16 model for that I assume?
just so you know, i got fluxmanias model early because they set their price to 100 buzz, i would have even paid 500 buzz for it. but you setting your model to 5000 means your not getting my buzz, or a lot of peoples buzz, because 5k is insane.
This is most agreeable.
Thank you for expressing interest in the model. I hope you consider enjoying it once it gets the normal non-early-access release <3
@6tZ Question: do you get real money in exchange for all that buzz? What's the purpose of it besides on site tools?
I want to train a Futanari/Newhalf LoRA for this model. Does it understand male genitalia? Do you have any model-specific advice for me? :)
It does, but it is still not as good at it as it is with female genitalia. I would suggest further fine-tuning in this niche with a large dataset, and lots of steps. Use prodigy and train for 100-300k steps.
Flux models are on the generator now, so this model is accessible for this current week. Give it a spin!
I tried to train lora with v2.4 and 3.0 and it is good, but when i tried to train with 4.0 fp8, my computer got hangs while training.
Is that because of the fp8 checkpoint?
I am using a RTX 3090. Thank you~
That should work just fine.
can't skip without saying sdxl is way better at this at this moment. faraway no chance .. flux just composition is better regards understand the prompt well , but render nudes or these tings , sdxl is way far better .
check out the stuff I've been making
Soup_Fusion the antichrist is going to emerge from your machine.
Please release fp16 models too. The fp8 and gguf versions fail to import in DrawThings.
Maybe in the future. These are fp8 versions.
this issue with FP8 E5M2 weights has been fixed in recent DT releases, assuming that's the cause with Fluxed Up
@Narz FP16 is out now for v5 and likely future versions as well.
help needed
Using xformers attention in VAE
VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
no CLIP/text encoder weights in checkpoint, the text encoder model will not be loaded.
Token indices sequence length is longer than the specified maximum sequence length for this model (225 > 77). Running this sequence through the model will result in indexing errors
Requested to load FluxClipModel_
Mmm...nsfw flux, here we go...
Please tell me which is the latest updated version for forge that I should use? If I download the standard download on the right arrow I download a file that on WebUI Forge gives me a lot of problems with deformed bodies. There are a thousand versions on this page, which one should I download?
Latest should be fine.
Deformed bodies would not come from forge, but rather from the models or settings or methods you are using.
You could also try alternatives like Fooocus, SwarmUI or ComfyUI.