Training epochs: 0
Training steps: 500
Learning rate: 1e-06
- Learning rate schedule: polynomial
- Warmup steps: 1000
Max grad norm: 2.0
Effective batch size: 2
- Micro-batch size: 2
- Gradient accumulation steps: 1
- Number of GPUs: 1
Gradient checkpointing: True
Prediction type: epsilon (extra parameters=['training_scheduler_timestep_spacing=trailing', 'inference_scheduler_timestep_spacing=trailing'])
Optimizer: adamw_bf16
Trainable parameter precision: Pure BF16
Caption dropout probability: 10.0%

Datasets

my-dataset-1024

Repeats: 63
Total number of images: 69
Total number of aspect buckets: 1
Resolution: 1.048576 megapixels
Cropped: False
Crop style: None
Crop aspect: None
Used for regularisation data: No

regularisation-data-1024px

Repeats: 0
Total number of images: 1000
Total number of aspect buckets: 1
Resolution: 1.048576 megapixels
Cropped: False
Crop style: None
Crop aspect: None
Used for regularisation data: Yes

Inference

python

import torch
from diffusers import DiffusionPipeline

model_id = 'FastFreddi/simpletuner-v1-full'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16

prompt = "a photograph of ohwx woman"
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
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(42),
    width=1024,
    height=1024,
    guidance_scale=4.2,
    guidance_rescale=0.0,
).images[0]
image.save("output.png", format="PNG")

Exponential Moving Average (EMA)

SimpleTuner generates a safetensors variant of the EMA weights and a pt file.

The safetensors file is intended to be used for inference, and the pt file is for continuing finetuning.

The EMA model may provide a more well-rounded result, but typically will feel undertrained compared to the full model as it is a running decayed average of the model weights.