<h3 id="introduction">Introduction</h3>Just another amateur remix of the Anima model.With this model, you can create many fascinating artworks, just like an amateur.<h3 id="required-components">Required Components</h3><ol><li>Text Encoder: <a target="_blank" rel="ugc" href="https://huggingface.co/circlestone-labs/Anima/resolve/main/split_files/text_encoders/qwen_3_06b_base.safetensors?download=true">https://huggingface.co/circlestone-labs/Anima/resolve/main/split_files/text_encoders/qwen_3_06b_base.safetensors?download=true</a></li></ol><ol><li>VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/circlestone-labs/Anima/resolve/main/split_files/vae/qwen_image_vae.safetensors?download=true">https://huggingface.co/circlestone-labs/Anima/resolve/main/split_files/vae/qwen_image_vae.safetensors?download=true</a></li></ol>Alternative VAE: <a rel="ugc" href="https://civarchive.com/models/2487530/qwenimagevaeliquid1087">QwenimageVAE_liquid1087</a> or <a rel="ugc" href="https://civarchive.com/models/2954301/qwenimagevaeliquid3844">Qwenimagevae_liquid3844</a><h3 id="recommended-settings">Recommended Settings</h3>Sampler: Euler A / ER SDE / dpmpp_2m_sde_gpu / res_multistepScheduler: Simple / Beta(57)Steps: 30-50 stepsCFG: 4-5If you're using <a target="_blank" rel="ugc" href="https://civarchive.com/models/2619830/turbo-for-anima-less-steps?modelVersionId=3039549">TURBO for ANIMA</a> or<a target="_blank" rel="ugc" href="https://civarchive.com/models/2466415/cosmos-predict25-2b-base-distilled-extracted-dmd2-lora"> Cosmos-Predict2.5-2B base distilled extracted DMD2 LoRA</a>, you might start off with 12 steps with CFG 1.0.<h3 id="known-issues-limitations">Known Issues / Limitations</h3><ol><li>Unless explicitly specified, version 1.0 is likely to produce characters with red eyes, especially under dark environment.</li><li>Stacking too many style LoRAs on top of this model might lead to over-saturation.</li><li>Malformed limbs or fingers on all versions - still figuring out how to fix. Adding 'bad anatomy' to the negative prompt seems better.</li><li>No up-to-date character knowledge - All character knowledge was acquired from the merged models. No plans for further improvement on this aspect.</li></ol><h3 id="tips-for-using-lora-with-2.9b">Tips for Using LoRA with 2.9b</h3>Latest version of Forge Neo supports Anima-2.9b and will do LoRA mapping automatically. For ComfyUI users, you might need to install <a rel="ugc" href="https://github.com/sparklingcoffee777/ComfyUI-Anima-2.9B-loraPatch">ComfyUI-Anima-2.9B-loraPatch</a> or <a rel="ugc" href="https://github.com/KitagawaCoding/ComfyUI-Anima-3.8B-loraPatch">ComfyUI-Anima-3.8B-loraPatch</a> for remapping LoRAs. Alternatively, you can use the LoRAs that is natively built for Anima-2.9b, for example, <a rel="ugc" href="https://civarchive.com/models/2857451/anima-29b-turbo-lora-remapped">Anima-2.9B Turbo LoRA Remapped</a>.<h3 id="license">License</h3>'Non-Commercial Use Only' under section 2(b):<a target="_blank" rel="ugc" href="https://huggingface.co/circlestone-labs/Anima/blob/main/LICENSE.md">https://huggingface.co/circlestone-labs/Anima/blob/main/LICENSE.md</a>

Amateur Remix

Experimental merge with Anima-2.9b<ul><li>Newer character knowledge</li></ul>Potential regressions<ul><li>Reduced color saturation/contrast</li><li>Compatibility issues</li><li>Slower inference speed</li><li>Furrification on well-known characters might not work</li></ul>To be tested<ul><li>Anatomy corrections</li><li>Pose</li><li>Consistency</li></ul>Merged models:Anima-2.9B + v1.4 + v1.5

Files with SHA256: 2c0781ac30bafe6c4e90dfe92fad7f6c1f6c1ccb32e5c037c8608ab706ee3a14