This is a blank / identity VAE created to solve a frustrating issue with certain checkpoint models in Forge, reForge and AUTOMATIC1111.I had to choose a "Base model", but this VAE works for SD 1.4, 1.5, 2.0, 2.1, SDXL, Pony, Lumina, Illustrious and NoobAI. Basically, anything that has "baked" or "built-in" VAE and does not require you to enable it. <h3 id="why-this-exists:-rpn6vt85f">Why this exists:</h3>When selecting a checkpoint model from the Checkpoints tab, some models include a line under:<code>Preferred VAE / Text encoder(s)</code> → <code>Built in</code>This causes issues like:<ul><li>Forge, reForge or A1111 not fully clearing the previous VAE</li><li>A misleading “built-in” VAE being shown when there’s none</li><li>The model continuing to use the last selected VAE, even when the field says “Built in”</li></ul>The result? Shimmering, JPEG-like artifacts, muddy contrast, and other unwanted effects or even crashes.<h3 id="what-this-does:-2wca5stff">What this does:</h3>This <code>.safetensors</code> file is a zeroed identity VAE that:<ul><li>Performs no encoding or decoding</li><li>Forces your UI to override the baked-in or ghosted VAE</li><li>Cleans up renders by guaranteeing a neutral, untouched output pipeline</li></ul>In layman's terms, this VAE does nothing. It is blank, null, empty. It forces your WebUI to switch VAE when switching checkpoints.<h3 id="how-to-use-it:-92dmzooce">How to use it:</h3><ol><li>Drop this into your <code>models/VAE/</code> folder</li><li>Select your model designated for "built in VAE" under the Checkpoints tab.</li><li>Click on the 🛠️ icon to "Edit metadata".</li><li>Under "Preferred VAE / Text encoder(s)", select "Built in.safetensors", then Save.</li></ol><h3 id="how-to-reproduce-the-issue-this-fixes:-z2trn8m2p">How to reproduce the issue this fixes:</h3>If you are not sure why a blank VAE is even needed, here is how to replicate the bug:<ol><li>Load a model that requires a VAE, such as an SDXL or FLUX checkpoint.</li><li>Go to the 🛠️ Edit Metadata, and under Preferred VAE, select a compatible VAE and Save.</li><li>Now, switch to a model like WAI-NSFW Illustrious or any other that has VAE already built-in.</li><li>In the metadata, choose "Built in" for the VAE and Save.</li><li>Switch between the two models again and look at the top center of the UI where Forge displays the active VAE/Text Encoder.</li></ol>You will notice that the previous VAE remains active, even though the new model is supposed to use none. This causes artifacts or improper rendering or even a crash until you force an override.That is exactly where this blank VAE comes in. Drop it in, select it manually, and you are safe.I hope this helps!You can find a handful of models with baked in/built-in VAE under "Suggested Resources"! ↴

Forge's "Built In" VAE

This VAE works with everything except FLUX, which requires a completely different VAE and text encoder setup. From SD 1.4, Pony, and SDXL, to Illustrious, NoobAI, and beyond, this blank VAE can be safely used to override built-in or buggy VAE behavior. I wish I could select more base models from the list, but that would probably just confuse people. Or maybe not. Honestly... I have no idea. :P

Files with SHA256: 816d0af04f76aed15e115699c769808480356e24e9f2f7657ec41d31f41e0553