Downloads · 30 days
0
Soushiki
Aotsuyu/Soushiki
Soushiki is a machine learning model from Aotsuyu. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as creativeml-openrail-m.
<center <h1 style="font-size:1.6rem;" <b Soushiki （葬式） </b </h1 </center The model some previously may have downloaded as 9-AAAAAAA is now renamed to SoushikiV0.9.
Access
Public
Updated Dec 11, 2023
Repo size
10.7 GB
Likes
7
Public
Repo makeup
Click a slice to open those files.
.safetensors8.5 GB · 100%
At a glance
- License
- creativeml-openrail-m
- Access
- Public
- Created
- Nov 21, 2023
- Updated
- Dec 11, 2023
- SHA
- 9d5795f4
Model card
From the Hugging Face model README
<center>
<h1 style="font-size:1.6rem;">
<b>
Soushiki （葬式）
</b>
</h1>
</center>
The model some previously may have downloaded as `9-AAAAAAA` is now renamed to SoushikiV0.9.
For 1.0 I very slightly slightly increased finger separation (it doesn't necessarily mean better hands, rather it's less likely that hands become an eldritch mess).
I bumped the vibrance a bit.
Do not hang up on the versions too much, 0.9 might be more to your liking.
previews
<img src="https://huggingface.co/Aotsuyu/Soushiki/resolve/main/img/xyz_grid-0043-2155415307.png" width="50%"></img> <img src="https://huggingface.co/Aotsuyu/Soushiki/resolve/main/img/xyz_grid-0044-384282104.png" width="50%"></img> <img src="https://huggingface.co/Aotsuyu/Soushiki/resolve/main/img/xyz_grid-0048-3671274917.png" width="50%"></img> <img src="https://huggingface.co/Aotsuyu/Soushiki/resolve/main/img/xyz_grid-0049-73335132.png" width="50%"></img> <img src="https://huggingface.co/Aotsuyu/Soushiki/resolve/main/img/xyz_grid-0050-667165639.png" width="50%"></img>
settings
Use whatever the fuck you want.
I keep my negative to low quality, worst quality, and usually use DPM++ 2M Karras or Restart samplers at 20-25 steps with 5-10 cfg.
kl-f8-anime2 or a variant for VAE.
ramblings
Yet another fairly shizo merge. I started off from Gishiki, adding elements I liked from CTD's lovely connectmix3 and Jemnite's amazing TangYuanV3.
This time I also utilized some artist LoRA and some diffLoRA to push the textures, contrast, details noise offset and such into directions I wanted.
The artstyle loras were mostly mine, but I got some off of Derrian and one anon.
I consider this less style neutral from what I usually go for, but I had no trouble using LoRAs with it.
This is a product of autistic XYZ comparison and abusing supermerger and I find it satisfying for now.
does this have new holos
No, I can't be arsed to make a finetune from scratch, maybe sometime in the future that will change.
Use this or whatever for holo advent https://huggingface.co/Aotsuyu/LoRA/blob/main/characters/%5Bchars%5D%20holoadventv5.safetensors
- License
- creativeml-openrail-m
- Created
- Nov 21, 2023
- Updated
- Dec 11, 2023