Sign In

Project Pygmalion

23

Updated: Jul 2, 2026

Download

1 variant available

bf16 SafeTensor

Jetlisted.safetensors_00001_.safetensors

BF16, good balance • 3.9 GB

Verified:

Type

Checkpoint

Stats

157

23

Reviews

Published

Jul 2, 2026

Base Model

Anima

Hash

AutoV2
4374B0463D
default creator card background decoration
Uploads - 262

262

Followers - 1844

1.8K

License:

Anima

The Anima Model is licensed by CircleStone Labs LLC. Copyright CircleStone Labs LLC. IN NO EVENT SHALL CIRCLESTONE LABS LLC BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.

Built on NVIDIA Cosmos

Project Pygmalion

A vibrant Illustrious merge focused on expression.

About the Model

This is a merge I put together using Illustrious checkpoints.
(Waikitty's Model Basket, Indexed, MaturaNero, and Hassaku)

My main goal with this project was to fix the "blank stare" issue I was getting with other models (suggested by @zekses). I wanted to create something where the characters actually look expressive and "alive," rather than just standing there.

Style-wise, it sits somewhere between anime and western comics. You can expect bold outlines, vibrant colors, and very clean composition without needing a massive negative prompt.

Important: Tagging

Use standard quality tags instead:

masterpiece, best quality, very aesthetic, absudres

My Settings

I’ve found this specific setup works best to get the look in the sample images (like the Miku example).

  • Sampler: DPM++ 3M SDE (or dpmpp_3m_sde_dynamic_eta in ComfyUI)

  • Scheduler: Kl_optimal

  • Steps: 35 (I prefer slightly higher steps for better detail)

  • CFG Scale: 5.0

  • Resolution: 896x1152 works great for portraits.

Prompting Tips

  • Framing: The tag cowboy_shot is great for getting that "knees-to-head" framing I use in the examples.

  • Style: Tags like comic style, bold outlines, and vibrant help bring out the model's specific look.