krea2 AKSIOM EXAPORN full precision
A balanced sub- model for realistic and semi-realistic NSFW image generation, developed as the core reference of the AKSIOM model family.
Training models and getting them ready for download takes a lot of work and time. Give me a thumbs-up if you use it. Thanks a million. Hope you like!
This Model is modified to gen highly precised human details. In a multistep development-process the semantics and the general Styles are aligned to a mostly representive NSFW Dynamic. We have a closed/locked (baked) Fingerprint.
Load Model via Unet-Loader. Use your Models for CLIP and VAE you best known/use.
LoRA´s are welcome, this Model will accept well trained Adaptations.
Sampler: Euler
Scheduler: beta
Steps: 8 (~9)
Resolution: (Krea2 typical used will accept)
Be carefully how your Prompts are build. This Model needs clear and precised Prompts.
Accepting typical SDXL-Prompts also, if you wanna use.
Additional Models Needed:
qwen3vl_4b_8fp_scaled.safetensors (CLIP)
qwen_image_vae.safetensors (VAE)
Resolution:
1536 x 2048
or common resolutions for Krea2 in general
Credits & Base Model
This model inherits its fundamental weighting from Krea-2 RAW and contains additional development work and weight modifications.
Please refer to the applicable license and usage conditions of the underlying model before redistributing or creating derivative works.
License & Derivative Work
krea2 AKSIOM exaporn 1 bf16 is a derivative work based on the Krea-2 architecture and underlying Krea-2 RAW model. Its development and distribution are therefore subject to the applicable terms, restrictions and requirements of the Krea-2 Community License.
Users downloading, using, modifying or redistributing this model are responsible for reviewing and complying with the Krea-2 Community License and its applicable guidelines. This includes any requirements concerning attribution, derivative works, redistribution and permitted use.
The additional weights and development work introduced for AKSIOM exaporn 1 do not replace, override or supersede the terms of the underlying Krea-2 Community License.
Parent models:
- krea2_aksiom_core_turbo_v1.safetensors
- krea2_aksiom_core_dev_detailer.savetensors
Derivative: krea2 AKSIOM exaporn 1 bf16
Final Notes
This release is intentionally called exaporn 1.
It is not meant to be the most specialized, most stylized or most extreme model in the AKSIOM family.
It is the model from which the family grows.
The development process behind Core Turbo Family was built around experimentation, comparison and controlled observation. The same approach will continue to guide future AKSIOM models.
If you experiment with the model, feedback, comparisons and unusual use cases are very welcome.
Have fun exploring it.
Description
full precision bf16 V01 exaporn
FAQ
Comments (8)
So, am I understanding correctly that this checkpoint is slightly more nsfw than the sannhet submodel or turbo foundation model? Int8 or fp8 coming soon?
@subterranian Yes, aksiom exaporn is explicit dev to nsfw in bf16. based on the Parent-Model aksiom core. sannhet is a child from core, like exaporn too. (architecture/Weighting).Other Variants for exaporn are in progress. needs time.
Sannhet is already an excellent model, but if it isn't explicit, what is its intended strength?
@subterranian sannhet 1 is dev for:
- dynamic
- Semirealistic -> Realistic
- NSFW
- Different Pussy-styles
- firm Breasts (diff sized)
- a balanced analog Look if its requested
- deep listening Prompt-accepting
Note: With unconventional or contradictory prompts, this model is prone to semantic misinterpretations.
You can read all in sannhet´s Modelcard. BF16 Version will released soon, and have the same Properties like the Int8 Version. With just a little more.....(secret).
nice job
int8_convrot - please
@Giovecivitai1 Sure, it's expected to cost 0,5 Money. Will you transfer the money as usual? 🤣🤣
But all joking aside, the version is coming—I'm not a factory, after all. So, thanks for the compliment. 😇
I converted this to nvfp4 and the quality is still exceptional. Nice work. Thank you for the full precision release.
@gurilagardnr Thank you - but remember, converting copy the Fingerprint too. Please note this. If you use it locally for you self its ok. But sharing is not allowed - sorry. ☝️ But i think you have no Problem to respect my work, or? 🫶