Right at the very beginning
Hi everyone, how are you? I'm doing well, thank you, and very busy creating Loras (I think I'll need to take a vacation later 🫠😵💫).
I don't know if you're aware, but in the last 3 months we've been bombarded with information and releases about a model called Anima, which seems very promising and a perfect candidate to replace Illustrious/NoobAI.
One of the main reasons for this is that the Anima model has several features that I will highlight here:
Main features:
- Better adherence to the prompt (including the ability to mix tags with natural language), incorporation of Qwen3 Embedding 6B (more information at this link - Qwen/Qwen3-Embedding-0.6B · Hugging Face)
- Reduction in model size from 6.4 GB (SDXL) to 3.89 GB
- Supports Qwen-Image VAE, resulting in more precise details such as eyes, face, and hands without necessarily requiring the use of ADetailer as in SDXL
- Increase in training resolution from 512px (preview version) to 1024px (base version)
- Adaptation and consistency of styles based on Gelbooru / Danbooru / DeviantArt (unfortunately, at the moment the model does not support E621 tags, meaning it is a bit more restrictive for characters furry, but you can incorporate it using Danbooru's tags)
Before I forget, I'll also be releasing a checkpoint (model) based on Anima soon. For now, it will be a more "refined" version, but I intend to include a larger base in the future, especially for furry characters (I plan to study this further).
Let's get down to business
In short, I will alternate between publishing and training Loras on Illustrious and Anima, so no one misses out (sorry Pony, you'll never be forgotten, but I'm short on time 😥).
Rest assured, I will do my best to maintain consistency and create Loras based on requests (I haven't forgotten about you 📌). Therefore, I will port all my Loras from Illustrious to Anima. Unfortunately, this will take a little more time, but I will do the same task as when I ported my Loras from Pony to Illustrious.
Don't worry, you can continue making requests. Whenever possible, I will complete them as quickly as possible. And in case of doubt, many people still don't know, but I have a Ko-Fi and I update the Lora statuses there based on requests. I avoid publishing updates here through articles because it's easier for me to get lost and also not receive notifications of updates.
A big hug to everyone. 😆👋