■This is an experimental ground for Anima LoRAs.
The LoRAs I create serve more of a role in stylistic tuning rather than reinforcing single concepts.
■I would like to share the possibilities of anima with everyone.
My wish is for many people to discover basemodels with potential and to see their possibilities unfold even further. I would be happy if I can help make that happen.
■The anima architecture has been heavily modified and retrained from scratch—almost like a radical overhaul—but the base architecture is Cosmos-Predict2-2B. To put it simply, you could compare it to a lightweight version of Flux 1.
I think it’s a great architecture capable of generating high-quality images despite its small size.
I have a very good impression of Anima when it comes to color vibrancy, lighting, and prompt adherence.
Its knowledge regarding NSFW concepts, characters, and styles is at least on par with SDXL. The overall feel of using it is also quite similar to SDXL.
Furthermore, it is lightweight and easy to train. I believe large-scale training is also a viable option.
Since we are provided with a base model that already contains a vast amount of knowledge, we rarely need to teach the model completely unknown concepts from scratch.
Basically, our work mostly comes down to adjusting styles or reinforcing minor concepts and characters. You can achieve great results without having to do any heavy lifting.
The developer has also ensured transparency by sharing the training settings, which means you are less likely to stumble over black-box issues.
Because of this, the burden of training is low, making it an option well worth nurturing together as a community.
■The LoRA has a stronger influence when used with shorter prompts.
Since longer prompts often carry their own inherent style, the LoRA will primarily take on a supporting role to refine and unify the overall image.
■https://github.com/gazingstars123/Anima-Standalone-Trainer
This is a great tool that allows for easy training, even on Windows. It has everything you need for LoRA training.
Even if you run into any issues, you should be able to get your LoRA training up and running by troubleshooting with ChatGPT or Gemini.
If you find the tool useful, you might want to consider making a donation to him. It allows the developer to focus more on development and deliver even better tools, which ultimately benefits you in the long run.
■If you have any questions, please feel free to ask!
日本語での質問も大丈夫ですのでご気軽にお声がけください～
Description
■This LoRA was trained on 21,080 handpicked anime images from Gelbooru. The dataset was carefully curated using both aesthetic scores and my own personal judgment to ensure high quality.
■It suppresses overly vivid or flat styles, instead producing images with a better sense of spatial depth, indirect lighting, and balanced contrast. This often gives the generations a much more cohesive, finished-artwork feel.
■A weight of 1.0 works perfectly fine, but if you notice the composition or details becoming unstable or breaking down, lowering the weight to 0.8 or 0.5 can help stabilize the image while still retaining the LoRA's benefits.
■You will notice its effects most prominently when using simple prompts or omitting standard "quality" tags. In those situations, where the prompt lacks specific style information and would normally result in a flat image, this LoRA steps in to guide the generation toward a more aesthetic result.
■On the other hand, if you are using long, detailed descriptions or heavy quality prompts, you likely already have a clear outcome in mind. In such cases, the LoRA won't completely dominate the image; instead, it will take on a supporting role. It helps by adding natural lighting and atmosphere, and if the model occasionally tries to generate an unintended style, the LoRA will course-correct it back in a more aesthetic direction.
■I'm not sure if anyone will want to, but you are more than welcome to use my LoRA in your merges and publish the resulting models.
However, please limit your merging to the "preview3" version.
This will help prevent "merge pollution"
Let's work together to build a transparent model ecosystem!
■I've also included backup LoRAs from different training steps, so feel free to experiment with them.
The 13,250-step and 14,250-step versions are looking quite good.
Due to Civitai's upload restrictions, the file extension has been changed to .bin.
Please rename the extension to .zip before extracting it.