Downloads · 30 days
20
7% of all-time downloads
sd3_vae_anime_ft
Disty0/sd3_vae_anime_ft
sd3_vae_anime_ft is a text-to-image model from Disty0. Use it when you need an image from a text prompt. It is set up for diffusers.
This VAE was finetuned on PNG only anime illustrations for 512 steps with MAE Loss. Used fp32 weights + fp16 mixed precision with learning rate 4e-6 and effective batch size of 16.
All-time downloads
300
Public
Parameters
83.8M
168 MB on disk
Likes
1
Public
Repo makeup
Click a slice to open those files.
.safetensors168 MB · 100%
At a glance
- Task
- Text-to-Image
- Library
- diffusers
- Access
- Public
- Created
- May 30, 2025
- Updated
- Jun 1, 2025
- SHA
- 4afe0572
Base models
Model card
From the Hugging Face model README
This VAE was finetuned on PNG only anime illustrations for 512 steps with MAE Loss.
Used fp32 weights + fp16 mixed precision with learning rate 4e-6 and effective batch size of 16.
This training was to test my VAE decoder training code and 512 step model turned out to be better than i expected and fixes the color shifting issues of the original SD3 VAE pretty well.
I stopped messing with SD3 after a while but i decided to release this VAE finetune instead of deleting it.
Original Image:
Original SD3 VAE:
Anime VAE Finetune:
- Task
- Text-to-Image
- Library
- diffusers
- Created
- May 30, 2025
- Updated
- Jun 1, 2025