Kirazuri (Anima)
Version 4.0 (Latest)
For in-depth details of training and tooling, see:
Training Details Summary
Trainer: diffusion-pipe
Training device: NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
Total training time: ~10 days
Total samples seen(unbatched steps): ~3,000,000
Training resolutions:
512^2
768^2
1024^2
1536^2
Stage 1
Samples seen(unbatched steps): ~2,000,000
Training time: ~125 hrs
Learning Rate: 6e-6
Learning Rate Scheduler: Cosine
LLM Adaptor Learning Rate: 8e-7
Precision: Mixed BF16
Optimizer: AdamW8bit with Kahan Summation
Weight Decay: 0.01
Timestep Sampling Strategy: Logit-Normal
Training Resolutions: 512^2, 768^2, 1024^2
Stage 2
Samples seen(unbatched steps): ~1,000,000
Training time: ~84 hrs
Learning Rate: 2e-6
Learning Rate Scheduler: Cosine
LLM Adaptor Learning Rate: 2e-7
Precision: Mixed BF16
Optimizer: AdamW8bit with Kahan Summation
Weight Decay: 0.01
Timestep Sampling Strategy: Logit-Normal
Training Resolutions: 512^2, 1024^2, 1536^2
Additional Features
Masked Training
Tag Dropout: 30% with protected first 8 tags
Tag Shuffle: Applied to last unprotected tags
Natural Language: Short and Long Caption variants
Changes from Kirazuri (Anima) v3.0
Dataset includes recently curated 2,450 images increasing total size from 42,608 to 45,058 images
Dataset cutoff now of 29/06/2026
Introduced Masked Training for images with simple backgrounds
Updated tags+caption variants structure
Recognitions
Thanks to Circlestone Labs for the Anima Preview base model.
Thanks to tdrussell of Circlestone Labs for the diffusion-pipe trainer.
Thanks to bluvoll for support using their fork of diffusion-pipe.
Thanks to narugo1992 and the deepghs team for open-sourcing various training sets, image processing tools, and models.
License
This model is released under the same license as the base model.
See the base model for details of the CircleStone Labs Non-Commercial License.