README.md

metadata

license: other
base_model: black-forest-labs/FLUX.1-dev
tags:
  - flux
  - flux-diffusers
  - text-to-image
  - diffusers
  - simpletuner
  - safe-for-work
  - lora
  - template:sd-lora
  - lycoris
inference: true
widget:
  - text: unconditional (blank prompt)
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_0_0.png
  - text: >-
      hckl_style, A collection of radiolarians floating in water. Various
      species with different shapes and structures.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_1_0.png
  - text: >-
      hckl_style, A large jellyfish with trailing tentacles. Smaller jellyfish
      surround it in the ocean.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_2_0.png
  - text: >-
      hckl_style, A diverse coral reef ecosystem. Various types of coral, sea
      anemones, and small fish.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_3_0.png
  - text: >-
      hckl_style, An assortment of diatoms. Different species showcasing their
      unique geometric structures.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_4_0.png
  - text: hckl_style, hamster
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_5_0.png
  - text: hckl_style, a hipster making a chair
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_6_0.png
  - text: >-
      hckl_style, A detailed evolutionary tree diagram showing the relationships
      between various species.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_7_0.png
  - text: >-
      hckl_style, A modern microbiology laboratory with researchers using
      advanced microscopes and equipment.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_8_0.png
  - text: >-
      hckl_style, A bustling city street with skyscrapers, cars, and
      pedestrians.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_9_0.png
  - text: >-
      hckl_style, An orbiting space station with astronauts conducting
      experiments in zero gravity.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_10_0.png
  - text: >-
      hckl_style, A person scrolling through a social media feed on a
      smartphone, surrounded by floating app icons.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_11_0.png
  - text: >-
      hckl_style, A energetic rock band performing on stage with a large crowd
      in the foreground.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_12_0.png
  - text: >-
      hckl_style, A visual representation of artificial intelligence, with
      interconnected nodes and data streams.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_13_0.png
  - text: a hamster, hckl_style
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_14_0.png

Flux-Ernst-Haeckel-LoKr-02

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

The main validation prompt used during training was:

a hamster, hckl_style

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

hckl_style, A collection of radiolarians floating in water. Various species with different shapes and structures.

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, A large jellyfish with trailing tentacles. Smaller jellyfish surround it in the ocean.

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, A diverse coral reef ecosystem. Various types of coral, sea anemones, and small fish.

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, An assortment of diatoms. Different species showcasing their unique geometric structures.

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, hamster

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, a hipster making a chair

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, A detailed evolutionary tree diagram showing the relationships between various species.

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, A modern microbiology laboratory with researchers using advanced microscopes and equipment.

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, A bustling city street with skyscrapers, cars, and pedestrians.

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, An orbiting space station with astronauts conducting experiments in zero gravity.

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, A person scrolling through a social media feed on a smartphone, surrounded by floating app icons.

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, A energetic rock band performing on stage with a large crowd in the foreground.

Negative Prompt

blurry, cropped, ugly

Prompt

hckl_style, A visual representation of artificial intelligence, with interconnected nodes and data streams.

Negative Prompt

blurry, cropped, ugly

Prompt

a hamster, hckl_style

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: 2
• Training steps: 7750
• Learning rate: 0.001
• Effective batch size: 1
  • Micro-batch size: 1
  • Gradient accumulation steps: 1
  • Number of GPUs: 1
• Prediction type: flow-matching
• 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

ernst-haeckel-flux-512

• Repeats: 10
• Total number of images: 81
• Total number of aspect buckets: 1
• Resolution: 0.262144 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None

ernst-haeckel-flux-1024

• Repeats: 10
• Total number of images: 81
• Total number of aspect buckets: 2
• Resolution: 1.048576 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None

ernst-haeckel-flux-512-crop

• Repeats: 10
• Total number of images: 81
• Total number of aspect buckets: 1
• Resolution: 0.262144 megapixels
• Cropped: True
• Crop style: random
• Crop aspect: square

ernst-haeckel-flux-1024-crop

• Repeats: 10
• Total number of images: 81
• 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 hamster, hckl_style"

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")