simpletuner-lora

This is a LyCORIS adapter derived from sd3/unknown-model.

The main validation prompt used during training was:

Create an image of a fabric design featuring a classic camouflage pattern in various shades of green, brown, and black, suitable for military clothing and outdoor gear.

Validation settings

CFG: 3.0
CFG Rescale: 0.0
Steps: 20
Sampler: None
Seed: 42
Resolutions: 1024x1024,1280x768

Note: The validation settings are not necessarily the same as the training settings.

You can find some example images in the following gallery:

Prompt
unconditional (blank prompt)

Negative Prompt
blurry, cropped, ugly

Prompt
Create an image of a fabric design featuring a classic camouflage pattern in various shades of green, brown, and black, suitable for military clothing and outdoor gear.

Negative Prompt
blurry, cropped, ugly

The text encoder was not trained. You may reuse the base model text encoder for inference.

Training settings

Training epochs: 8
Training steps: 3500
Learning rate: 0.0001
Max grad norm: 0.01
Effective batch size: 1
- Micro-batch size: 1
- Gradient accumulation steps: 1
- Number of GPUs: 1
Prediction type: flow-matching
Rescaled betas zero SNR: False
Optimizer: adamw_bf16
Precision: Pure BF16
Quantised: No
Xformers: Not used
LyCORIS Config:

{
    "algo": "lokr",
    "multiplier": 1.0,
    "linear_dim": 10000,
    "linear_alpha": 1,
    "factor": 16,
    "apply_preset": {
        "target_module": [
            "Attention",
            "FeedForward"
        ],
        "module_algo_map": {
            "Attention": {
                "factor": 16
            },
            "FeedForward": {
                "factor": 8
            }
        }
    }
}

Datasets

GPTwebd3

Repeats: 0
Total number of images: 409
Total number of aspect buckets: 1
Resolution: 1.0 megapixels
Cropped: True
Crop style: center
Crop aspect: square
Used for regularisation data: No

Inference

import torch
from diffusers import DiffusionPipeline
from lycoris import create_lycoris_from_weights

model_id = '/root/autodl-tmp/stable-diffusion-3-medium-diffusers'
adapter_id = 'pytorch_lora_weights.safetensors' # you will have to download this manually
lora_scale = 1.0
wrapper, _ = create_lycoris_from_weights(lora_scale, adapter_id, pipeline.transformer)
wrapper.merge_to()

prompt = "Create an image of a fabric design featuring a classic camouflage pattern in various shades of green, brown, and black, suitable for military clothing and outdoor gear."
negative_prompt = 'blurry, cropped, ugly'
pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu')
image = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    num_inference_steps=20,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(1641421826),
    width=1024,
    height=1024,
    guidance_scale=3.0,
).images[0]
image.save("output.png", format="PNG")