■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.There is no inherent superiority among the LoRAs; they simply serve different purposes. You can find the description for each LoRA in its respective tab.■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.■<a target="_blank" rel="ugc" href="https://github.com/gazingstars123/Anima-Standalone-Trainer">https://github.com/gazingstars123/Anima-Standalone-Trainer</a>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～

anima_lora_lab

Model OverviewI created this LoRA using 5,080 aesthetic 2D anime-style images. It was trained for over 130 epochs and 23000+ steps. With an effective batch size of 32, the training is quite deep. To ensure stability, the main file is a merge of LoRAs from several different epochs.Using a weight between 0.5 and 0.8 might be easier to handle, as it reduces the chance of artifacts and image degradation.Style Tendencies &amp; EffectsThis LoRA actively suppresses the realistic/2.5D "AI-like" look, consistently forcing a pure 2D style. It also fine-tunes spatial lighting and color balance. As a result, it strongly enhances an organic vibe, making the outputs look as if they were drawn by a real human artist.While the 2D aesthetic remains highly consistent, the specific stylistic direction will be determined by the tags and seed you use.Comparison with 2D_aesthetic_lora_base-v1.0Because this LoRA uses a larger dataset than 2D_aesthetic_lora_base-v1.0, it offers greater stylistic diversity and overall versatility. However, it is not necessarily a strict upgrade.A smaller dataset actually has a higher probability of consistently hitting one specific style.Tip: If you combine both LoRAs together, you can achieve the best of both worlds—the vast diversity of the large dataset combined with the striking, consistent aesthetics of the smaller one.Backup Versions IncludedI am also sharing several backup LoRAs saved at different training stages. The style tendencies shift depending on the step count, so I recommend testing them out to find the one that fits your personal preference. Additionally, since these backups are unmerged, their LoRA influence applies much more directly to the generation.

Files with SHA256: 9824549be94c3d123be7fc3b8a2ba48876324facb99d3240ac22f15b28af6d93