Model description for Flux-Sketch-Smudge-LoRA
Image Processing Parameters
Parameter |
Value |
Parameter |
Value |
LR Scheduler |
constant |
Noise Offset |
0.03 |
Optimizer |
AdamW |
Multires Noise Discount |
0.1 |
Network Dim |
64 |
Multires Noise Iterations |
10 |
Network Alpha |
32 |
Repeat & Steps |
22 & 3290 |
Epoch |
18 |
Save Every N Epochs |
1 |
Labeling: florence2-en(natural language & English)
Total Images Used for Training : 26 [ 14 bit raw ]
Best Dimensions & Inference
Dimensions |
Aspect Ratio |
Recommendation |
1280 x 832 |
3:2 |
Best |
1024 x 1024 |
1:1 |
Default |
Inference Range
- Recommended Inference Steps: 30–35
Setting Up
import torch
from pipelines import DiffusionPipeline
base_model = "black-forest-labs/FLUX.1-dev"
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=torch.bfloat16)
lora_repo = "strangerzonehf/Flux-Sketch-Smudge-LoRA"
trigger_word = "Sketch Smudge"
pipe.load_lora_weights(lora_repo)
device = torch.device("cuda")
pipe.to(device)
Trigger words
You should use Sketch Smudge
to trigger the image generation.
Download model
Weights for this model are available in Safetensors format.
Download them in the Files & versions tab.