Raehoshi Illust XL
an enhanced iteration built upon the Illustrious XL model. It aims to elevate the visual style by addressing some of the limitations in the original, such as oversaturation and artifact noise. While these issues are not entirely eliminated, noticeable improvements have been made. The goal is to deliver a more polished, balanced output while staying true to the strengths of the base model.
Why Early Access?
Early access helps keep the project going. I don’t have my own GPU, so all training is done through rented cloud GPUs and that gets pretty expensive. By getting early access, you’re directly supporting the development of my models and helping me keep improving them. If you'd like to support me further, you can also buy me a coffee on Ko-fi! Every bit of help means a lot and keeps the future updates coming.
Recommended setting
Positive prompt :
masterpiece, best quality, very aesthetic, absurdresNegative prompt :
bad quality, worst quality, jpeg artifacts, sketch, bad anatomy, signature, watermarkSteps : 25+
CFG : 5-7
Sampler : euler a or dpm++2m karras (euler for vpred)
Standard resolution :
832 x 1216, 1216 x 832, 1152 x 896, 896 x 1152, 1344 x 768, 768 x 1344, 1024 x 1024High resolution :
1024 x 1536, 896 x 1536, 1536 x 1024, 1536 x 896Hires.fix Setting:
Upscaler : 4x Foolhardy Remacri
Hires step : 10-15
Denoise : 0.1-0.3
Special Tags
Quality Tags:
masterpiece
best quality
good quality
average quality
bad quality
worst quality
Rating Tags:
safe
sensitive
nsfw
nsfw, explicit
Aesthetic Tags:
very aesthetic
aesthetic
displeasing
very displeasing
Training Details
The model was developed using a two-stage fine-tuning process. In Stage 1, new series and characters were introduced into the model. Stage 2 focused on fixing issues and enhancing the overall style for improved output.
Stage 1
Dataset : v1-31k, v2-37k, v3-34k, v4-60k, v5_v5.1-18k, v6-15k, v7-39k, v8-41k, v9-30k, v10-30k with multi resolution
Hardware : 2x A100 80gb, v3, v4, v5, v5.1-2x H100 80gb, v7,v8, v9, v10-RTX PRO 6000
Batch size : 32
Gradient accumulation steps : 2
Learning rate : 6e-6
Text encoder : 3e-6
Epoch : 15
Stage 2
Dataset : v1-2.5k, v2 and v3-2.3k, v4-2.5k, v5-2k, v5.1-1.8k, v6-1.5k, v7-1.7k, v7.1,v8-4.1k, v9-1.9k, v10-2.4k
Hardware : 1x A100 80gb, v7_v7.1,v8, v9, v10-RTX PRO 6000
Batch size : 48
Gradient accumulation steps : 1
Learning rate : 3e-6, v5.1-2.5e-6
Text encoder : disable
Epoch : 15
List of New Series/Characters Trained:
Zenless Zone Zero
Wuthering Waves
Honkai: Star Rail
Genshin Impact
Arknights: Endfield
Umamusume
Azur Lane
Arknights
Fate/GO
Dandadan
Make heroine ga oo sugiru
Kusuriya no Hotorigoto
Hololive from justice and dev is
Indie Vtuber Dooby, Yuuki Sakuna, Nimi Nightmare, and S***
100 girlfriends who really love you
Haite kudasai takamine-san
Alina clover
Nikke: bready and little mermaid
Kpop Demon Hunters
Full character list are available article here
For character trait details prompts, please refer to the Danbooru site for accurate tags and references.
License
Special thanks to Joe for supporting my works
Special thanks to Juno for supporting my works and help me with early tester
Description
Fixed Catastrophic Forgetting: Resolved issues where certain characters from previous data were lost.
Improved Stability: Enhanced overall model consistency and output quality.
Note: If you encounter any issues with specific characters, please let me know!
FAQ
Comments (8)
boss, will aemeath from wuwa be properly baked into the model with high fidelity? love your work
Hi! Actually, Aemeth is already included in the model. However, since v9 was trained last month right when she was released, there wasn't much fan art available at the time. I'm planning to add more data for her in the next update to improve the fidelity. Thanks for the support
@Raelina really appreciate it, she is the latest favorite from the wuwa community, btw how do i make the model to generate older characters, it seem without positive/negative prompting for age/mature appearance it default to a young girl appearnce, which i hope to generate adult women
@asdasdsad I think there might be a slight misunderstanding regarding the 'older characters' mentioned in the changelog. What I meant was characters from previous training versions (like v8) that became harder to generate in v9. I'm sorry for the confusion. I’ll make sure to clarify the changelog.
If you're looking to generate a more adult or mature appearance, I recommend adding tags like mature female or mature woman to your prompt. Thank you
It's marvelous models!!
I’ve been wondering about this for a while, are you merging Lora with the model to train new characters?
Thank you! I actually use full/native training rather than merging Lora to add new characters to the model. This helps maintain higher quality and consistency. You can find more specifics about my process in the Training Details section on the model page.
@Raelina Oh, native training?
I’m pretty sure native training is a pretty time-consuming method, do all model trainers these days use native training to fine-tune their models?
@Rayla You're right, native training is indeed very time-consuming and expensive. That’s why I offer early access to help cover the costs, as I mainly rent GPUs from RunPod for the training.
As for other trainers, it varies. A good way to tell is by checking the 'Checkpoint Type' in the model details. Generally, a 'Checkpoint Trained' implies full finetuning, while a 'Checkpoint Merge' is often created by combining Loras or other models. That’s just my perspective, though every trainer has their own preferred workflow