simpletuner-lora

simpletuner-lora

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

The main validation prompt used during training was:

A simplistic, hand-drawn illustration of an elephant. the elephant is depicted in a walking pose, with its trunk raised slightly. the drawing is done in black ink on a white background. the elephant's posture and the positioning of its legs suggest movement. the style is minimalistic, with clean lines and a lack of intricate details. the lighting appears to be coming from the top left, casting a shadow on the right side of the elephant.

Validation settings

• CFG: 7.5
• CFG Rescale: 0.0
• Steps: 35
• Sampler: FlowMatchEulerDiscreteScheduler
• Seed: 42
• Resolution: 512x512
• 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: 6

• Training steps: 4000

• Learning rate: 0.0001

  • Learning rate schedule: cosine
  • Warmup steps: 100

• Max grad norm: 0.01

• Effective batch size: 16

  • Micro-batch size: 4
  • Gradient accumulation steps: 4
  • Number of GPUs: 1

• Gradient checkpointing: True

• Prediction type: flow-matching (extra parameters=['shift=3'])

• Optimizer: adamw_bf16

• Trainable parameter precision: Pure BF16

• Caption dropout probability: 10.0%

• LoRA Rank: 128

• LoRA Alpha: None

• LoRA Dropout: 0.1

• LoRA initialisation style: default

Datasets

pacs

• Repeats: 0
• Total number of images: 9980
• Total number of aspect buckets: 1
• Resolution: 1.0 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 = 'Cha-Imaa/simpletuner-lora'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "A simplistic, hand-drawn illustration of an elephant. the elephant is depicted in a walking pose, with its trunk raised slightly. the drawing is done in black ink on a white background. the elephant's posture and the positioning of its legs suggest movement. the style is minimalistic, with clean lines and a lack of intricate details. the lighting appears to be coming from the top left, casting a shadow on the right side of the elephant."
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=35,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(42),
    width=512,
    height=512,
    guidance_scale=7.5,
).images[0]
image.save("output.png", format="PNG")