JuanGris-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: 3.0
CFG Rescale: 0.0
Steps: 20
Sampler: FlowMatchEulerDiscreteScheduler
Seed: 42
Resolution: 896x1280
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
In the style of a Juan Gris cubist painting, a coffee pot and two cups sit on a café table. A folded newspaper and half-eaten croissant rest nearby.

Negative Prompt
blurry, cropped, ugly

Prompt
In the style of a Juan Gris cubist painting, a mandolin rests on a wooden stand next to sheet music. A glass of wine and pipe lie on the side table.

Negative Prompt
blurry, cropped, ugly

Prompt
In the style of a Juan Gris cubist painting, an ink bottle stands next to a fountain pen and open journal. Several sealed letters and a wax seal rest alongside them.

Negative Prompt
blurry, cropped, ugly

Prompt
In the style of a Juan Gris cubist painting, a brass telescope points skyward through a domed window. Star charts and astronomical instruments lie on a curved desk.

Negative Prompt
blurry, cropped, ugly

Prompt
In the style of a Juan Gris cubist painting, a vintage sewing machine sits centered on a wooden table. Spools of thread, scissors, and fabric patterns are scattered around it.

Negative Prompt
blurry, cropped, ugly

Prompt
In the style of a Juan Gris cubist painting, a hamster sits in its glass habitat next to its exercise wheel. A water bottle and food dish are visible inside.

Negative Prompt
blurry, cropped, ugly

Prompt
In the style of a Juan Gris cubist painting, a classic Coca-Cola bottle stands on a table next to an empty glass with ice cubes. Condensation beads on both surfaces.

Negative Prompt
blurry, cropped, ugly

Prompt
In the style of a Juan Gris cubist painting, a Range Rover SUV is parked in front of a modern garage. Its distinctive grille and headlights face forward.

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: 22
Training steps: 9500
Learning rate: 0.0001
- Learning rate schedule: constant
- 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=['flux_schedule_auto_shift', 'shift=0.0', 'flux_guidance_mode=constant', 'flux_guidance_value=1.0', 'flux_beta_schedule_alpha=4.0', 'flux_beta_schedule_beta=2.0', '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

juangris-512

Repeats: 10
Total number of images: 23
Total number of aspect buckets: 7
Resolution: 0.262144 megapixels
Cropped: False
Crop style: None
Crop aspect: None
Used for regularisation data: No

juangris-1024

Repeats: 10
Total number of images: 23
Total number of aspect buckets: 12
Resolution: 1.048576 megapixels
Cropped: False
Crop style: None
Crop aspect: None
Used for regularisation data: No

juangris-512-crop

Repeats: 10
Total number of images: 23
Total number of aspect buckets: 1
Resolution: 0.262144 megapixels
Cropped: True
Crop style: random
Crop aspect: square
Used for regularisation data: No

juangris-1024-crop

Repeats: 10
Total number of images: 23
Total number of aspect buckets: 1
Resolution: 1.048576 megapixels
Cropped: True
Crop style: random
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/JuanGris-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=896,
    height=1280,
    guidance_scale=3.0,
).images[0]
image.save("output.png", format="PNG")

davidrd123
/

JuanGris-Flux-LoKr