|
--- |
|
language: |
|
- en |
|
tags: |
|
- stable-diffusion |
|
- text-to-image |
|
license: unknown |
|
datasets: |
|
- Danbooru2021 |
|
inference: true |
|
|
|
--- |
|
|
|
## Repo Overview |
|
|
|
A version of Waifu Diffusion v1.4 Float32 Booru 110k model and VAE from Waifu Diffusion v1.4 converted to work with Diffusers library. |
|
|
|
 |
|
^ Quality/Style test |
|
|
|
 |
|
^ Unnatural resolution test |
|
|
|
 |
|
^ Landscape test |
|
|
|
 |
|
^ Extremely long resolution test |
|
|
|
## Model Description |
|
|
|
The current model has been fine-tuned on a Stable Diffusion 2.1 model with 110k anime-styled images using a technique known as aspect ratio bucketing. |
|
|
|
That allows Waifu Diffusion v1.4 to handle different resolutions much better than its previous models. |
|
|
|
## Source |
|
|
|
WD v1.4 Model:https://huggingface.co/hakurei/waifu-diffusion-v1-4/blob/9fa4a42a9c4a0948472fa909e6c1a39be0dda699/models/wd-1-4-float32-booru-110k.ckpt |
|
|
|
WD v1.4 VAE: https://huggingface.co/hakurei/waifu-diffusion-v1-4/blob/main/vae/kl-f8-anime2.ckpt |
