This LoRA is trained on a synthetic dataset of health bars, one full one on the top, one empty one on the bottom. However, I failed to train it to maintain that symmetry/shape. Thus it's recommended to just use img2img with some rough sketch as input. Keywords:<ul><li>health bar</li><li>left_aligned, centered, right_aligned</li><li>left_tilted, right_tilted</li><li>power_meter</li><li>no_segments, small_segments, large_segments</li><li>pixel_art, minimalist, sleek, ornate, cartoon, photo_realistic, abstract, vintage, surreal, fantasy, stylized</li><li>red, orange, yellow, lime, green, cyan, light_blue, blue, indigo, violet, purple, magenta, pink, brown, gray, black, white, gold, silver, teal</li></ul>

Files with SHA256: 8020a43e22d4a8a036383778e1fe8003bdcf40499d53623f5a0dd65b5653c31b