README.md

---
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
---
<b>It's important! Read it!</b><br>
The model is still in the training phase. This is not the final version. This version of the model may contain errors in anatomy and text, and it may also contain artifacts. I tried to fix them, and this is the best result so far compared to what was achieved previously.<br>

<b>You can support me directly on Boosty - https://boosty.to/sg_161222</b>

<h2>Using the Model</h2>
<b>Compact Version</b> The Compact version already includes VAE, Clip-L, and T5XXL.
  <ul>
      <li>For ComfyUI, place the model in the following path: ComfyUI_windows_portable\ComfyUI\models\checkpoints</li>
      <li>For Forge, place the model in the following path: webui_forge_cu121_torch231\webui\models\Stable-diffusion</li>
  </ul>

<b>Transformer Version (Unet)</b> The Transformer version requires you to have Flux VAE, Clip-L, and T5XXL.
  <ul>
      <li>For ComfyUI:</li>
          <ul>
              <li>Place the VAE in the following path: ComfyUI_windows_portable\ComfyUI\models\vae</li>
              <li>Place Clip-L and T5XXL in the following path: ComfyUI_windows_portable\ComfyUI\models\clip</li>
              <li>Place the Transformer (Unet) in the following path: ComfyUI_windows_portable\ComfyUI\models\unet</li>
          </ul>
  </ul>

Then you can use the Basic Workflow by moving the image into the ComfyUI workspace.

<h2>Optimal Generation Settings</h2>

<b>Sampling Method and Schedule Type:</b> Euler Beta<br>
<b>Sampling Steps:</b> 25+<br>
<b>Distilled CFG Scale:</b> 3.5<br>
<b>CFG Scale:</b> 1.0

<h2>Limitations of RealFlux 1.0b Dev</h2>
    <ul>
        <li>This version of the model does not have a variety of styles and is intended for realism (styles will be added with each new version of the model).</li>
        <li>There may be issues with anatomy and text.</li>
    </ul>