simpletuner-fulltraining

This is a full rank finetune derived from stabilityai/stable-diffusion-3-medium-diffusers.

The main validation prompt used during training was:

a three fourth perspective portrait view of a young woman with messy blonde hair and light purple eyes, looking at viewer with a closed mouth smile

Validation settings

• CFG: 3.0
• 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:

Prompt

unconditional (blank prompt)

Negative Prompt

blurry, cropped, ugly

Prompt

a three fourth perspective portrait view of a young woman with messy blonde hair and light purple eyes, looking at viewer with a closed mouth smile

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: 2
• Training steps: 24001
• Learning rate: 1e-05
  • Learning rate schedule: polynomial
  • Warmup steps: 100
• 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: 0.05%

Datasets

my-dataset-1024

• Repeats: 10
• Total number of images: 460
• Total number of aspect buckets: 8
• Resolution: 1.048576 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

my-dataset-crop-1024

• Repeats: 10
• Total number of images: 455
• Total number of aspect buckets: 1
• Resolution: 1.048576 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

Inference

import torch
from diffusers import DiffusionPipeline

model_id = 'RobinHCL/simpletuner-fulltraining'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16

prompt = "a three fourth perspective portrait view of a young woman with messy blonde hair and light purple eyes, looking at viewer with a closed mouth smile"
negative_prompt = 'blurry, cropped, ugly'

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=3.0,
).images[0]

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