How to use matteomarjanovic/flatsketcher with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-schnell", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("matteomarjanovic/flatsketcher") prompt = "The technical flat sketch shows a sleeveless top with a sophisticated architectural design. The front view features a dramatic V-neckline that plunges deeply, created by overlapping panels. The bodice maintains a clean, fitted silhouette through the torso. At the waist, there's a distinct horizontal band or waistband with a moderate width, likely 2-3 inches, that sits at the natural waistline. The most distinctive detail is an asymmetrical draped panel that extends from the left side of the waistband, falling below the hip line in a diagonal sweep. This panel appears to be constructed from the same fabric as the main body and creates an intentional gathered effect where it connects to the waistband. The back view would mirror the clean lines of the front bodice, maintaining the sleeveless armholes with clean finished edges. The V-neckline on the back would be higher than the front for structural integrity. The waistband continues around the entire circumference of the garment with consistent width. The asymmetrical draped panel visible from the front would be partially visible from the back view, showing how it wraps around the left side of the garment. Construction details would include darts or princess seams for fitting, though these are minimally visible to maintain the clean aesthetic. The garment likely closes with an invisible zipper at the center back seam above the waistband for ease of dressing. In the style of FLTSKC" image = pipe(prompt).images[0]
