Download
1 variant available
bf16 SafeTensor
SnakebitesHoopLipRings_v1_EarlyAccess_Unstable_krea2.safetensors
BF16, good balance • 382.14 MB
Verified: 2 days ago
This version is intended for Early Access and is considered unstable. I will refine the training and release the next version soon.

3800 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
2.5K0 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9K
Following the release of my Septum Hoop LoKR, I received a request to train a separate concept for snakebites. I put together this v1 Early Access version to get it into your hands, but please read the notes below before using it!
Due to having to rely heavily on a dataset of synthetic/AI-edited images, this first version is unstable and somewhat rigid compared to my Septum model. It absolutely works, but it requires some prompting finesse. I am releasing this as an Early Access V1 for those who want to start generating right away, while I continue cooking V2.
⚠️ Capabilities & Current Limitations
Trigger Word: SnakebitesHoopLipRings
Art Styles: V1 strongly prefers Realism and CGI / 3D. It currently struggles heavily with applying to non-human characters.
Customization: Currently limited compared to the Septum LoKR. It generates fairly large average standard hoops and can handle larger thicker hoops, and reliably supports only three colors: Black, Silver, and Gold.
Combo Friendly: The good news! I have thoroughly tested this alongside my SeptumHoopNoseRing LoKR. The two concepts combine beautifully and produce highly coherent imagery together. Stacking them also retains the side-effect of enhancing the coherency of ear gauge plugs when prompted.
⚙️ Recommended Usage & Settings
Weight/Strength: This is where the instability shows the most. The strength requirement is all over the place depending on your prompt, base model, and seed. You will need to experiment anywhere from 0.45 up to 1.4.
Samplers: Like the septum model, tested best on Euler Simple or ddim with a ddim_uniform scheduler.
🛠️ Under the Hood & Plans for V2
For the technical artists and trainers:
This was trained on 116 images for 20 epochs using the Fizgig trainer using the ultra fast preset.
☕ Support & Future Plans
Like my previous Krea2 release, this V1 model is temporarily locked behind paid BUZZ access to help recoup the hardware and compute costs associated with my AMD training setup. As soon as this model hits its break-even point, I will flip the switch and make it 100% free for everyone forever.
If you want to support my work outside of Civitai, you can check out my [Patreon Link]. (Note: I do not keep exclusive models paywalled there; it's strictly an alternative tip jar for those who wish to support my art and dev work).
Commissions: I'm currently open for custom LoRA / LoKR training commissions (private use or expedited public release). Feel free to reach out via DM! (Note: I do not accept commissions to train on IRL persons).
Thank you for the support, and have fun experimenting with V1!

