stable-diffusion-3.5-medium-HC

stable-diffusion-3.5-medium-HC

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

The main validation prompt used during training was:

Designed by Hyundai, 3/4 front view, red Palisade SUV parked in a misty forest clearing, early morning fog and golden sun rays filtering through pine trees, dew drops on side mirror, Hyundai emblem shining, lush green plants, dramatic backlight, reflections of the woods on glass, ultra-realistic design, cool color grading, natural sunlight, macro details, premium feel, elegant and powerful stance.

Validation settings

• CFG: 7.5
• CFG Rescale: 0.0
• Steps: 20
• Sampler: FlowMatchEulerDiscreteScheduler
• Seed: 42
• Resolution: 1024x1024
• Skip-layer guidance:

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: 4

• Training steps: 3000

• Learning rate: 0.0001

  • Learning rate schedule: sine
  • Warmup steps: 500

• Max grad value: 2.0

• 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=3'])

• Optimizer: adamw_bf16

• Trainable parameter precision: Pure BF16

• Base model precision: no_change

• Caption dropout probability: 0.0%

• LoRA Rank: 256

• LoRA Alpha: 256.0

• LoRA Dropout: 0.1

• LoRA initialisation style: default

• LoRA mode: Standard

Datasets

test1_data_set

• Repeats: 5
• Total number of images: 101
• Total number of aspect buckets: 1
• Resolution: 1.048576 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

Inference

import torch
from diffusers import DiffusionPipeline

model_id = 'stabilityai/stable-diffusion-3.5-medium'
adapter_id = 'mingyu-oo/stable-diffusion-3.5-medium-HC'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "Designed by Hyundai, 3/4 front view, red Palisade SUV parked in a misty forest clearing, early morning fog and golden sun rays filtering through pine trees, dew drops on side mirror, Hyundai emblem shining, lush green plants, dramatic backlight, reflections of the woods on glass, ultra-realistic design, cool color grading, natural sunlight, macro details, premium feel, elegant and powerful stance."
negative_prompt = 'blurry, cropped, ugly'

## Optional: quantise the model to save on vram.
## Note: The model was not quantised during training, so it is not necessary to quantise it 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
model_output = 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(42),
    width=1024,
    height=1024,
    guidance_scale=7.5,
).images[0]

model_output.save("output.png", format="PNG")