flux-juggalos-lokr

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

The main validation prompt used during training was:

a juggalo holding a sign that says keepin it real

Validation settings

CFG: 3.4
CFG Rescale: 0.0
Steps: 25
Sampler: None
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 juggalo performing miracles at a festival

Negative Prompt
blurry, cropped, ugly

Prompt
a juggalo man and woman in a close-up portrait smiling and happy

Negative Prompt
blurry, cropped, ugly

Prompt
a juggalo family portrait

Negative Prompt
blurry, cropped, ugly

Prompt
a juggalo teacher giving a lesson to a classroom of juggalo students

Negative Prompt
blurry, cropped, ugly

Prompt
a toddler juggalo criminal mugshot 1998 olan mills studio photography

Negative Prompt
blurry, cropped, ugly

Prompt
a baby juggalo welder working on the titanic, 1912, sepia

Negative Prompt
blurry, cropped, ugly

Prompt
a baby juggalo news anchor on the insane clown news network wearing glasses

Negative Prompt
blurry, cropped, ugly

Prompt
a juggalo holding a sign that says keepin it real

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: 8
Training steps: 2500
Learning rate: 5e-06
Effective batch size: 12
- Micro-batch size: 6
- Gradient accumulation steps: 1
- Number of GPUs: 2
Prediction type: flow-matching
Rescaled betas zero SNR: False
Optimizer: optimi-lion
Precision: bf16
Quantised: No
Xformers: Not used
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

juggalos-aspect-bucket

Repeats: 10
Total number of images: ~26
Total number of aspect buckets: 7
Resolution: 0.262144 megapixels
Cropped: False
Crop style: None
Crop aspect: None

juggalos-crop

Repeats: 10
Total number of images: ~22
Total number of aspect buckets: 1
Resolution: 0.262144 megapixels
Cropped: True
Crop style: random
Crop aspect: square

juggalos-aspect-bucket-1mp

Repeats: 10
Total number of images: ~26
Total number of aspect buckets: 8
Resolution: 1.048576 megapixels
Cropped: False
Crop style: None
Crop aspect: None

juggalos-crop-1mp

Repeats: 10
Total number of images: ~20
Total number of aspect buckets: 1
Resolution: 1.048576 megapixels
Cropped: True
Crop style: random
Crop aspect: square

Inference

import torch
from diffusers import DiffusionPipeline
from lycoris import create_lycoris_from_weights

model_id = 'black-forest-labs/FLUX.1-dev'
adapter_id = 'pytorch_lora_weights.safetensors' # you will have to download this manually
lora_scale = 1.0
wrapper, _ = create_lycoris_from_weights(lora_scale, adapter_id, pipeline.transformer)
wrapper.merge_to()

prompt = "a juggalo holding a sign that says keepin it real"

pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu')
image = pipeline(
    prompt=prompt,
    num_inference_steps=25,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(1641421826),
    width=1024,
    height=1024,
    guidance_scale=3.4,
).images[0]
image.save("output.png", format="PNG")

RareConcepts
/

Flux.1-dev-LoKr-Juggalos