wizard-s-vintage-etchings

Wizard's Vintage Etchings

<Gallery />

(CivitAI)

Model description

<p><strong><span style="color:#fa5252">Etching</span></strong><span style="color:#fa5252"> </span><strong><span style="color:#fa5252">art style</span></strong> LoRA that produces mostly monochrome images. </p><p>Use <code>etching,</code> in your prompt to trigger the style. <br /></p><p><strong>Contrast Guide</strong><br /><code>dark background</code> for more shadows <br /><code>bright background </code>for more contrast <br /><code>plain background </code>for more isolated subject</p><p><span style="color:#868e96">Play around with your </span><code>CFG</code><span style="color:#868e96"> settings, monochrome images usually handles a higher CFG than usual. A higher CFG gives more contrast and a lower CFG will make it softer.</span></p><p></p><p><strong>LoRA Strength Guide</strong><br /><span style="color:#868e96">Strength 0.5 = Less style, more tidy and clean, better coherence<br />Strength 0.80 = More style, mid coherent images<br />Strength 1.0 = Very stylized, more abstract, more impressionistic </span><br /><br /><br />Other settings: </p><p><code>Checkpoint: PixelWave 09</code><br /><code>Sampler: DPM++ 2m / DPM++ SDE</code> <br /><code>Scheduler: Align Your Steps (AYS)</code><br /><code>CFG: 2.5</code><br /><br />Notice:<br />Trained on around 40 high-res prints at 1536px max.res. <br />Most of the training images are from early 20th century scandinavia (Anders Zorn), so you might run in to some specific bias from that time-era/location.</p>

Trigger words

You should use Etching, dark background, bright and plain background, pencil sketch, oil painting to trigger the image generation.

Download model

Weights for this model are available in Safetensors format.

Download them in the Files & versions tab.

Use it with the 🧨 diffusers library

from diffusers import AutoPipelineForText2Image
import torch

device = "cuda" if torch.cuda.is_available() else "cpu"

pipeline = AutoPipelineForText2Image.from_pretrained('stabilityai/stable-diffusion-xl-base-1.0', torch_dtype=torch.float16).to(device)
pipeline.load_lora_weights('WizWhite/wizard-s-vintage-etchings', weight_name='Wizards_Etchings-000008.safetensors')
image = pipeline('`Etching`, `dark background`, `bright and plain background`, `pencil sketch`, `oil painting`').images[0]

For more details, including weighting, merging and fusing LoRAs, check the documentation on loading LoRAs in diffusers