Flux-Actors-Face-Inset_Cig-Cards-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: 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

In the style of an 4ct0rs cigarette card, green-skinned ambassador in Victorian diplomatic uniform with ceremonial tentacle sash. HIGNETT'S CIGARETTES, 'AMBASSADOR ZXYLK'

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of an 4ct0rs cigarette card, distinguished werewolf dean in tweed jacket with silver-rimmed spectacles. HIGNETT'S CIGARETTES, 'DEAN WOLFGANG'

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of an 4ct0rs cigarette card, copper-featured android soprano in shimmering 1920s evening gown. HIGNETT'S CIGARETTES, 'MADAME MECHANICA'

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of an 4ct0rs cigarette card, pale-skinned magistrate in black judicial robes with crimson cravat. HIGNETT'S CIGARETTES, 'JUSTICE BLOODWORTH'

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of an 4ct0rs cigarette card, diminutive sovereign wearing dewdrop crown and spider-silk robes. HIGNETT'S CIGARETTES, 'QUEEN THISTLEDOWN'

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of an 4ct0rs cigarette card, decorated aquatic commander in mother-of-pearl armor with coral medals. HIGNETT'S CIGARETTES, 'GENERAL WAVEBREAKER'

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of an 4ct0rs cigarette card, lion-headed academic in scholar's robes and ancient headdress. HIGNETT'S CIGARETTES, 'DOCTOR RIDDLEWHISPER'

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of an 4ct0rs cigarette card, shrewd green-skinned financier in pinstripe suit with golden monocle. HIGNETT'S CIGARETTES, 'DIRECTOR GOLDGRASP'

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of an 4ct0rs cigarette card, romantic horseman wearing flowing poet's shirt and laurel crown. HIGNETT'S CIGARETTES, 'LORD HOOFINGTON'

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of an 4ct0rs cigarette card, translucent civic leader wearing mayoral chains and Victorian suit. HIGNETT'S CIGARETTES, 'MAYOR PHANTOMSWORTH'

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of an 4ct0rs cigarette card, scaled culinary master in pristine white uniform with flame-proof medals. HIGNETT'S CIGARETTES, 'CHEF FLAMESCALE'

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of an 4ct0rs cigarette card, temporal enforcement officer in quantum-tailored suit with chronometer badge. HIGNETT'S CIGARETTES, 'INSPECTOR CHRONOS'

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: 0
• Training steps: 600
• Learning rate: 0.0008
• Max grad norm: 2.0
• Effective batch size: 4
  • Micro-batch size: 4
  • Gradient accumulation steps: 1
  • Number of GPUs: 1
• Prediction type: flow-matching (flux parameters=['flux_guidance_value=1.0'])
• Rescaled betas zero SNR: False
• Optimizer: adamw_bf16
• Precision: Pure BF16
• Quantised: Yes: int8-quanto
• 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

actors-natural-512

• Repeats: 17
• Total number of images: 68
• Total number of aspect buckets: 1
• Resolution: 0.262144 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

actors-natural-768

• Repeats: 17
• Total number of images: 68
• Total number of aspect buckets: 1
• Resolution: 0.589824 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

actors-natural-1024

• Repeats: 5
• Total number of images: 68
• Total number of aspect buckets: 1
• Resolution: 1.048576 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

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

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=20,
    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=3.0,
).images[0]
image.save("output.png", format="PNG")