Training epochs: 1
Training steps: 4000
Learning rate: 8e-05
- Learning rate schedule: polynomial
- Warmup steps: 100
Max grad value: 0.1
Effective batch size: 2
- Micro-batch size: 2
- Gradient accumulation steps: 1
- Number of GPUs: 1
Gradient checkpointing: True
Prediction type: flow-matching (extra parameters=['shift=3', 'flux_guidance_mode=constant', 'flux_guidance_value=1.0', 'flow_matching_loss=compatible', 'flux_lora_target=all'])
Optimizer: adamw_bf16
Trainable parameter precision: Pure BF16
Base model precision: int8-quanto
Caption dropout probability: 3.0%
LoRA Rank: 64
LoRA Alpha: None
LoRA Dropout: 0.1
LoRA initialisation style: default

Datasets

klimt-256

Repeats: 10
Total number of images: 94
Total number of aspect buckets: 1
Resolution: 0.065536 megapixels
Cropped: False
Crop style: None
Crop aspect: None
Used for regularisation data: No

klimt-crop-256

Repeats: 10
Total number of images: 94
Total number of aspect buckets: 1
Resolution: 0.065536 megapixels
Cropped: True
Crop style: center
Crop aspect: square
Used for regularisation data: No

klimt-512

Repeats: 4
Total number of images: 94
Total number of aspect buckets: 13
Resolution: 0.262144 megapixels
Cropped: False
Crop style: None
Crop aspect: None
Used for regularisation data: No

klimt-crop-512

Repeats: 4
Total number of images: 94
Total number of aspect buckets: 1
Resolution: 0.262144 megapixels
Cropped: True
Crop style: center
Crop aspect: square
Used for regularisation data: No

klimt-768

Repeats: 3
Total number of images: 92
Total number of aspect buckets: 17
Resolution: 0.589824 megapixels
Cropped: False
Crop style: None
Crop aspect: None
Used for regularisation data: No

klimt-crop-768

Repeats: 3
Total number of images: 88
Total number of aspect buckets: 1
Resolution: 0.589824 megapixels
Cropped: True
Crop style: center
Crop aspect: square
Used for regularisation data: No

klimt-1024

Repeats: 2
Total number of images: 86
Total number of aspect buckets: 18
Resolution: 1.048576 megapixels
Cropped: False
Crop style: None
Crop aspect: None
Used for regularisation data: No

klimt-crop-1024

Repeats: 2
Total number of images: 84
Total number of aspect buckets: 1
Resolution: 1.048576 megapixels
Cropped: True
Crop style: center
Crop aspect: square
Used for regularisation data: No

klimt-1440

Repeats: 0
Total number of images: 64
Total number of aspect buckets: 9
Resolution: 2.0736 megapixels
Cropped: False
Crop style: None
Crop aspect: None
Used for regularisation data: No

klimt-crop-1440

Repeats: 0
Total number of images: 75
Total number of aspect buckets: 1
Resolution: 2.0736 megapixels
Cropped: True
Crop style: center
Crop aspect: square
Used for regularisation data: No

Inference

python

import torch
from diffusers import DiffusionPipeline

model_id = 'black-forest-labs/FLUX.1-dev'
adapter_id = 'bluezabu/klimt-lora'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "An astronaut is riding a horse through the jungles of Thailand."


## 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
model_output = pipeline(
    prompt=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(1337),
    width=1024,
    height=1024,
    guidance_scale=3.0,
).images[0]

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