OdilonRedon-QuarterCrops-Flux-LoKr

This is a LyCORIS adapter derived from black-forest-labs/FLUX.1-dev.

No validation prompt was used during training.

None

Validation settings

• CFG: 4.0
• CFG Rescale: 0.0
• Steps: 20
• Sampler: FlowMatchEulerDiscreteScheduler
• Seed: 42
• Resolution: 768x1280
• 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

RedonStyle painting, a figure with an ethereal golden halo, wearing dark flowing robes with gold accents, surrounded by vibrant orange and red flowers against a deep blue atmospheric background

Negative Prompt

blurry, cropped, ugly

Prompt

RedonStyle painting, a mystical arrangement of wildflowers in a blue vase, mixing poppies and strange botanical specimens in vivid reds and purples, set against a background of muted earth tones and shadowy forms

Negative Prompt

blurry, cropped, ugly

Prompt

RedonStyle painting, a figure in profile wearing a medieval blue headpiece with white trim, enveloped by a soft golden glow, with abstract floral patterns and scattered stars in the dark background

Negative Prompt

blurry, cropped, ugly

Prompt

RedonStyle painting, butterflies with luminous wings hovering over a garden of otherworldly flowers, their petals in shades of orange and blue, beneath a sky filled with swirling golden light

Negative Prompt

blurry, cropped, ugly

Prompt

RedonStyle painting, a winged horse emerging from a mystical forest, surrounded by floating spheres of light and exotic red flowers, with a deep blue-green sky filled with celestial patterns

Negative Prompt

blurry, cropped, ugly

Prompt

RedonStyle painting, an imagined landscape where giant flowers bloom beneath medieval arches, strange fruits glow with inner light, and spiritual figures float among clouds of gold and violet

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: 3

• Training steps: 5500

• Learning rate: 0.0004

  • Learning rate schedule: polynomial
  • Warmup steps: 100

• Max grad norm: 0.1

• Effective batch size: 3

  • Micro-batch size: 3
  • 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=4.0', 'flow_matching_loss=compatible'])

• Optimizer: adamw_bf16

• Trainable parameter precision: Pure BF16

• Caption dropout probability: 10.0%

• SageAttention: Enabled inference

LyCORIS Config:

{
    "algo": "lokr",
    "multiplier": 1.0,
    "linear_dim": 10000,
    "linear_alpha": 1,
    "factor": 16,
    "apply_preset": {
        "target_module": [
            "Attention",
            "FeedForward"
        ],
        "module_algo_map": {
            "Attention": {
                "factor": 16
            },
            "FeedForward": {
                "factor": 8
            }
        }
    }
}

Datasets

or-256

• Repeats: 10
• Total number of images: 37
• Total number of aspect buckets: 2
• Resolution: 0.065536 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

or-crop-256

• Repeats: 10
• Total number of images: 37
• Total number of aspect buckets: 1
• Resolution: 0.065536 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

or-512

• Repeats: 10
• Total number of images: 37
• Total number of aspect buckets: 4
• Resolution: 0.262144 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

or-crop-512

• Repeats: 10
• Total number of images: 37
• Total number of aspect buckets: 1
• Resolution: 0.262144 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

or-768

• Repeats: 10
• Total number of images: 37
• Total number of aspect buckets: 9
• Resolution: 0.589824 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

or-crop-768

• Repeats: 10
• Total number of images: 37
• Total number of aspect buckets: 1
• Resolution: 0.589824 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

or-1024

• Repeats: 10
• Total number of images: 37
• Total number of aspect buckets: 11
• Resolution: 1.048576 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

or-crop-1024

• Repeats: 10
• Total number of images: 37
• Total number of aspect buckets: 1
• Resolution: 1.048576 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

or-1440

• Repeats: 10
• Total number of images: 37
• Total number of aspect buckets: 13
• Resolution: 2.0736 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

or-crop-1440

• Repeats: 10
• Total number of images: 37
• Total number of aspect buckets: 1
• Resolution: 2.0736 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

Inference

import torch
from diffusers import DiffusionPipeline
from lycoris import create_lycoris_from_weights


def download_adapter(repo_id: str):
    import os
    from huggingface_hub import hf_hub_download
    adapter_filename = "pytorch_lora_weights.safetensors"
    cache_dir = os.environ.get('HF_PATH', os.path.expanduser('~/.cache/huggingface/hub/models'))
    cleaned_adapter_path = repo_id.replace("/", "_").replace("\\", "_").replace(":", "_")
    path_to_adapter = os.path.join(cache_dir, cleaned_adapter_path)
    path_to_adapter_file = os.path.join(path_to_adapter, adapter_filename)
    os.makedirs(path_to_adapter, exist_ok=True)
    hf_hub_download(
        repo_id=repo_id, filename=adapter_filename, local_dir=path_to_adapter
    )

    return path_to_adapter_file
    
model_id = 'black-forest-labs/FLUX.1-dev'
adapter_repo_id = 'davidrd123/OdilonRedon-QuarterCrops-Flux-LoKr'
adapter_filename = 'pytorch_lora_weights.safetensors'
adapter_file_path = download_adapter(repo_id=adapter_repo_id)
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
lora_scale = 1.0
wrapper, _ = create_lycoris_from_weights(lora_scale, adapter_file_path, pipeline.transformer)
wrapper.merge_to()

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
image = 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(42),
    width=768,
    height=1280,
    guidance_scale=4.0,
).images[0]
image.save("output.png", format="PNG")