Update: The "Kreanima" version is an extract of the <a target="_blank" rel="ugc" href="https://civarchive.com/models/2800474/kreanima">Kreanima</a> Model, meaning this is Lianima4Krea together with some layer of Dmerge.This will allow people using Kreanima without going for a BF16 or GGUF version (since i still didn't took the time to make an INT8 Convrot version and i feel a LoRA is better suited for Krea2 due to the size ^^;)<h3 id="lianima-for-krea2">Lianima for Krea2</h3>So, i dropped the dataset for <a target="_blank" rel="ugc" href="https://civarchive.com/models/2727072/lianima-personal-anima-base">Lianima </a>as-is in the on-site trainer and did 10 epochs of training on Krea2 just to experiment. In the end, i am quite happy with the result which will allow me to test some Krea2 LoRA on top of my own style! 💕<s>Since i can't run it locally yet, i need to make it available to test it on-site (i don't have a membership anymore, i can't test from the trainer).</s> (That's no longer the case, yeah!)NB: since it is an ANIMA dataset, it's tag based, sorry for the NLP lovers, still not my forte 😅NB2: The LoRA has a strong bias toward female character and Liadora base outfit (naked shoulder dress, black clothes).Usage (personal preference):<ul><li>Steps: 10</li><li>CFG: 1.2</li><li>Sampler: ER_SDE</li><li>Scheduler: Beta (or smoothstep for sd.cpp)</li></ul>All showcases pictures (except the first few tests) are generated using <a target="_blank" rel="ugc" href="https://github.com/leejet/stable-diffusion.cpp">stable-diffusion.cpp</a> (which now support Krea2, yeah! Local gen is back!)<ul><li>VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/blob/main/split_files/vae/wan_2.1_vae.safetensors">https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/blob/main/split_files/vae/wan_2.1_vae.safetensors</a></li><li>TE: <a target="_blank" rel="ugc" href="https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct-GGUF/blob/main/Qwen3VL-4B-Instruct-Q8_0.gguf">https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct-GGUF/blob/main/Qwen3VL-4B-Instruct-Q8_0.gguf</a></li><li>Turbo: <a target="_blank" rel="ugc" href="https://huggingface.co/realrebelai/KREA-2_GGUFs/blob/main/TURBO/Krea-2-Turbo-Q5_K_S.gguf">https://huggingface.co/realrebelai/KREA-2_GGUFs/blob/main/TURBO/Krea-2-Turbo-Q5_K_S.gguf</a></li></ul>The model should also work well on Raw, in this case, go for CFG 3.5 and 30 steps from my first tests.Version difference<ul><li>Vanilla is an average of epoch 6-7-8-9 of the pure training result. It suffers from censorship obviously even if some NSFW pictures were present and the "classic" LoRA filter did not help (i tried several and even fiddled a bit with numbers). But <a target="_blank" rel="ugc" href="https://civarchive.com/models/2728644/krea-2-mystic-xxx">this LoRA </a>by @alcaitiff did! 💜</li><li>Uncensored is an ugly merge of the <a target="_blank" rel="ugc" href="https://civarchive.com/models/2728644/krea-2-mystic-xxx">Mystic XXX LoRA</a> at strength 0.8, both using standard layers addition and SVD. Since neither was exactly what i wanted, i re-did a merge between those. I did this merge for personal usage as i would rather only load one LoRA and use base Turbo. You can either use Vanilla on an uncensored model or use the source Lora which is great! 😍 (I may do a model merge and then a quantization to have LoRA-included version locally, not quite sure it would be useful to provide it here).</li></ul>Thanks for reading! 💜

Lianima 4 Krea2

Posted on Civitai, upload found elsewhere are FORBIDDEN and done without the author permission.

Files with SHA256: 0bb04bc0b692eeb18c413feae8d17b5348cfb3d89e0c0f8b9a72e1c474e4a34f