import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", dtype=torch.bfloat16, device_map="cuda")
pipe.load_lora_weights("Migga/flux-scratch-lora-rk16-random")
prompt = "High contrast scratch defect on dark glass display, thin linear scratch, occasional diagonal orientation, sharp edges, isolated single defect, reflective glossy surface with subtle metallic sheen, fine texture on smooth surface, close-up industrial inspection photo, uniform lighting with faint glow, minimal dark background, minimal noise, shallow depth of field"
image = pipe(prompt).images[0]flux-scratch-lora-rk16-random
A Flux LoRA trained on a local computer with Fluxgym
- Prompt
- High contrast scratch defect on dark glass display, thin linear scratch, occasional diagonal orientation, sharp edges, isolated single defect, reflective glossy surface with subtle metallic sheen, fine texture on smooth surface, close-up industrial inspection photo, uniform lighting with faint glow, minimal dark background, minimal noise, shallow depth of field
Trigger words
You should use High contrast scratch defect on dark glass display, thin linear scratch, occasional diagonal orientation, sharp edges, isolated single defect, reflective glossy surface with subtle metallic sheen, fine texture on smooth surface, close-up industrial inspection photo, uniform lighting with faint glow, minimal dark background, minimal noise, shallow depth of field to trigger the image generation.
Download model and use it with ComfyUI, AUTOMATIC1111, SD.Next, Invoke AI, Forge, etc.
Weights for this model are available in Safetensors format.
- Downloads last month
- 20
Model tree for Migga/flux-scratch-lora-rk16-random
Base model
black-forest-labs/FLUX.1-dev