File size: 1,245 Bytes
bded544 5f08b65 bded544 2f66141 a2b33b3 2f66141 3ab8372 cfa32aa adf6aab 7634562 adf6aab |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 |
---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co./black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
---
# Verus Vision — Model for Realistic Image Generation
# Overview
**Verus Vision** is a fine-tuned model designed for generating high-quality, realistic images. Built on the **Flux Dev (de-distill)** model with 12 billion parameters.
<hr>
# How to Use Transformer Version
**For ComfyUI:**
* Place the **AE** in the following path: **ComfyUI_windows_portable\ComfyUI\models\vae**
* Place **Clip-L** and **T5XXL** in the following path: **ComfyUI_windows_portable\ComfyUI\models\clip**
* Place the **Transformer** in the following path: **ComfyUI_windows_portable\ComfyUI\models\unet** or **ComfyUI_windows_portable\ComfyUI\models\diffusion_models**
<hr>
# Optimal Generation Settings
* Sampling Method and Schedule Type: Euler Beta
* Sampling Steps: 25+
* CFG Scale: 3.5
> [!WARNING]
> **Warning:** Do not use a negative prompt as it can lead to the generation of images with artifacts and distortions.
<hr>
# Limitations
* This version of the model does not have a variety of styles and is intended for realism and photorealism.
* There may be issues with anatomy and text. |