mini-mix-lora-very-small-4

mini-mix-lora-very-small-4

This is a standard PEFT LoRA derived from stabilityai/stable-diffusion-3.5-large.

The main validation prompt used during training was:

mixqwerpoi The bright and clean kitchen area. The walls are finished with white paint, and the ceiling is also neatly finished with white paint. The floor is finished with light wood in a herringbone pattern, further emphasizing the bright feel of the entire space.The main color is white, silver (from the appliances and fixtures) is used as a secondary color, and warm wood tones are utilized as an accent color. White makes the space look wide and clean, silver adds a sleek modern touch, and the wood elements add warmth and naturalness. White cabinetry lines the kitchen, offering a clean and modern feel. A built-in microwave and dishwasher keep the space tidy, and a sleek faucet is installed beneath a window fitted with silver blinds. A white coffee machine sits on the counter, accompanied by wooden cutting boards that add a natural accent. Overall, it shows a minimal and sophisticated style. A bright, clean, and comfortable atmosphere is felt throughout the kitchen.

Validation settings

• CFG: 3.0
• CFG Rescale: 0.0
• Steps: 25
• Sampler: FlowMatchEulerDiscreteScheduler
• Seed: 42
• Resolutions: 1024x1024, 1024x896, 1280x768
• Skip-layer guidance: skip_guidance_layers=[7, 8, 9],

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

You can find some example images in the following gallery:

<Gallery />

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

Training settings

• Training epochs: 20

• Training steps: 10000

• Learning rate: 1e-05

  • Learning rate schedule: polynomial
  • Warmup steps: 500

• Max grad norm: 0.01

• Effective batch size: 1

  • Micro-batch size: 1
  • Gradient accumulation steps: 1
  • Number of GPUs: 1

• Gradient checkpointing: True

• Prediction type: flow-matching (extra parameters=['shift=1.0'])

• Optimizer: adamw_bf16

• Trainable parameter precision: Pure BF16

• Caption dropout probability: 40.0%

• LoRA Rank: 128

• LoRA Alpha: None

• LoRA Dropout: 0.1

• LoRA initialisation style: default

Datasets

mini-mix-lora-very-small-4

• Repeats: 30
• Total number of images: 16
• Total number of aspect buckets: 1
• Resolution: 1.048576 megapixels
• Cropped: True
• Crop style: random
• Crop aspect: preserve
• Used for regularisation data: No

Inference

import torch
from diffusers import DiffusionPipeline

model_id = 'stabilityai/stable-diffusion-3.5-large'
adapter_id = 'daehuncho/mini-mix-lora-very-small-4'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "mixqwerpoi The bright and clean kitchen area. The walls are finished with white paint, and the ceiling is also neatly finished with white paint. The floor is finished with light wood in a herringbone pattern, further emphasizing the bright feel of the entire space.The main color is white, silver (from the appliances and fixtures) is used as a secondary color, and warm wood tones are utilized as an accent color. White makes the space look wide and clean, silver adds a sleek modern touch, and the wood elements add warmth and naturalness. White cabinetry lines the kitchen, offering a clean and modern feel. A built-in microwave and dishwasher keep the space tidy, and a sleek faucet is installed beneath a window fitted with silver blinds. A white coffee machine sits on the counter, accompanied by wooden cutting boards that add a natural accent. Overall, it shows a minimal and sophisticated style. A bright, clean, and comfortable atmosphere is felt throughout the kitchen."
negative_prompt = 'blurry, cropped, ugly'

## Optional: quantise the model to save on vram.
## Note: The model was quantised during training, and so it is recommended to do the same during inference time.
from optimum.quanto import quantize, freeze, qint8
quantize(pipeline.transformer, weights=qint8)
freeze(pipeline.transformer)
    
pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu') # the pipeline is already in its target precision level
image = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    num_inference_steps=25,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(42),
    width=1024,
    height=1024,
    guidance_scale=3.0,
    skip_guidance_layers=[7, 8, 9],
).images[0]
image.save("output.png", format="PNG")