Introduction
This is a LyCORIS model for SDXL Anime checkpoints, enhancing the ability of the model to generate pictures of bdsm.
Usage
Trigger words: bdsm, rope
Artist prompts and style tags are recommended.
Recommended checkpoints:
Animagine XL (V2.0 only)
Resolution
This model is trained for resolutions from ARB 1024x1024 with minimum resolution 256 and maximum resolution 1536.
Training Details
Hardware: 1x RTX 4090
Num Train Images: 1496
Num Reg Images: 121
Batch Size: 4
Grad Accumulation Step: 4
Total Epoch: 6
Total Steps: 3488
Optimizer: AdamW8bit
Learning Rate: 8e-5 for UNet (train UNet only)
LR Scheduler: cosine_with_restarts
Caption Tag Dropout Rate: 0.1
Multires Noise Iterations: 6
Multires Noise Discount: 0.3
Resolution: 1024x1024
Min Bucket Resolution: 256
Max Bucket Resolution: 1536
Mixed Precision: FP16
Description
Improve the compatibility of Animagine XL
FAQ
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.