metadata

license: creativeml-openrail-m
base_model: stabilityai/stable-diffusion-xl-base-1.0
tags:
  - sdxl
  - sdxl-diffusers
  - text-to-image
  - image-to-image
  - diffusers
  - simpletuner
  - not-for-all-audiences
  - lora
  - controlnet
  - template:sd-lora
  - standard
pipeline_tag: text-to-image
inference: true
widget:
  - text: unconditional (blank prompt)
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_0_0.png
  - text: A photo-realistic image of a cat
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_1_0.png
  - text: prompt not found (2)
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_2_0.png
  - text: prompt not found (3)
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_3_0.png

simpletuner-controlnet-sdxl-lora-test

This is a ControlNet PEFT LoHa derived from stabilityai/stable-diffusion-xl-base-1.0.

The main validation prompt used during training was:

A photo-realistic image of a cat

Validation settings

CFG: 4.2
CFG Rescale: 0.0
Steps: 20
Sampler: ddim
Seed: 42
Resolution: 1024x1024

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 photo-realistic image of a cat

Negative Prompt
blurry, cropped, ugly

Prompt
prompt not found (2)

Negative Prompt
blurry, cropped, ugly

Prompt
prompt not found (3)

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: 49
Training steps: 400
Learning rate: 0.0001
- Learning rate schedule: constant
- Warmup steps: 0
Max grad value: 2.0
Effective batch size: 3
- Micro-batch size: 1
- Gradient accumulation steps: 1
- Number of GPUs: 3
Gradient checkpointing: True
Prediction type: epsilon (extra parameters=['training_scheduler_timestep_spacing=trailing', 'inference_scheduler_timestep_spacing=trailing'])
Optimizer: bnb-lion8bit
Trainable parameter precision: Pure BF16
Base model precision: no_change
Caption dropout probability: 0.1%
LoRA Rank: 128
LoRA Alpha: 128.0
LoRA Dropout: 0.1
LoRA initialisation style: default

Datasets

antelope-data

Repeats: 0
Total number of images: ~24
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 = 'stabilityai/stable-diffusion-xl-base-1.0'
adapter_id = 'bghira/simpletuner-controlnet-sdxl-lora-test'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "A photo-realistic image of a cat"
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.unet, weights=qint8)
#freeze(pipeline.unet)
    
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=4.2,
    guidance_rescale=0.0,
).images[0]

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