<p>A pixel art LoRA for general-purpose game assets such as character sprites, creatures, items/equipment, backgrounds, scenery, and icons.</p><p></p><h3 id="how-to-use">How to use</h3><p>You can use the default ERNIE-Image-Turbo workflow from ComfyUI, and no prompt enhancer needed. The sample images also have workflows.</p><p></p><h3 id="how-to-get-pixel-perfect-images">How to get pixel-perfect images</h3><p>Downscale by a factor of 4. So 512x512 images should downscale to 128x128, 1024x1024 to 256x256, and so on. Using k-centroid with something like <a target="_blank" rel="ugc" href="https://github.com/KohakuBlueleaf/PixelOE">PixelOE</a> works well.</p><p></p><h3 id="does-this-lora-work-with-ernie-image-base">Does this LoRA work with ERNIE-Image base?</h3><p>Yes, but I don't recommend it. <strong>The LoRA is meant to be used with the turbo model.</strong> For some reason, outputs with the base model are very bad. The colors are way too bright or saturated, and there are more issues with anatomy. Maybe there's a problem with my settings.</p><p></p><h3 id="notes-and-issues">Notes &amp; Issues</h3><p>There are still some issues with certain prompts with the ERNIE turbo model.</p><ul><li><p>The model tends to make characters face forward or in a 3/4 angle even if your prompt has a different view. This might just be a limit of the turbo model, though.</p></li><li><p>If prompting for sprites, make sure to include "white background" somewhere, otherwise you'll sometimes get a detailed background.</p></li><li><p>Since I trained this on a 4x upscaled pixel art dataset, if you want smaller sprites, just prompt for copies of a sprite in a 2x2 or 4x4 grid (see the sample images).</p></li><li><p>The dataset this LoRA was trained on contains 512x512, 768x768, and 1024x1024 images, but you can change the resolution and still get decent images.</p></li></ul><p></p>

Pixel Art Assets - General

<p>Trained on ERNIE-Image base. Use the LoRA with ERNIE-Image-Turbo. Example images were generated with the Q6_K GGUF, but you can swap out the model files.</p>

Files with SHA256: 2a2d582dfe8f01e8b586634bfce2f9667192fdcf8c80219abee72382f65f075e