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
  - standard
inference: true
widget:
  - text: unconditional (blank prompt)
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_0_0.png
  - text: >-
      This vibrant photograph captures a woman standing on a wooden deck, gazing
      out at a stunning, clear, turquoise ocean below. The woman is a tall, slim
      blonde with long, straight hair flowing down her back. She wears a bright
      turquoise bikini that contrasts with the vivid blue water and highlights
      her tanned skin. The wooden deck is sturdy and well-built, with a railing
      that she leans on for support. The deck is positioned on a cliff that
      overlooks a breathtaking seascape. The water is a gradient of vibrant
      turquoise and deep blue, with a sandy beach visible at the bottom of the
      image. The beach is pristine white, with gentle waves lapping against it.
      Three sailboats are anchored in the water, adding a sense of scale and
      perspective. The cliffs are lush with green vegetation, contrasting with
      the blue water and sky. The sky is a clear, bright blue, with no clouds.
      The overall mood is one of serene, tranquil relaxation, with the bright
      colors and clear, sharp details adding to the vividness and clarity of the
      scene.
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_1_0.png

deephouse-st-2911

This is a standard PEFT LoRA derived from black-forest-labs/FLUX.1-dev.

The main validation prompt used during training was:

This vibrant photograph captures a woman standing on a wooden deck, gazing out at a stunning, clear, turquoise ocean below. The woman is a tall, slim blonde with long, straight hair flowing down her back. She wears a bright turquoise bikini that contrasts with the vivid blue water and highlights her tanned skin. The wooden deck is sturdy and well-built, with a railing that she leans on for support. The deck is positioned on a cliff that overlooks a breathtaking seascape. The water is a gradient of vibrant turquoise and deep blue, with a sandy beach visible at the bottom of the image. The beach is pristine white, with gentle waves lapping against it. Three sailboats are anchored in the water, adding a sense of scale and perspective. The cliffs are lush with green vegetation, contrasting with the blue water and sky. The sky is a clear, bright blue, with no clouds. The overall mood is one of serene, tranquil relaxation, with the bright colors and clear, sharp details adding to the vividness and clarity of the scene.

Validation settings

• CFG: 3.0
• CFG Rescale: 0.0
• Steps: 20
• Sampler: FlowMatchEulerDiscreteScheduler
• Seed: 42
• Resolution: 1344x768
• 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

a women laughing with short hair

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: 1

• Training steps: 500

• Learning rate: 0.0004

  • Learning rate schedule: polynomial
  • Warmup steps: 100

• Max grad norm: 2.0

• Effective batch size: 1

  • Micro-batch size: 1
  • 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=1.0', 'flow_matching_loss=compatible', 'flux_lora_target=all'])

• Optimizer: adamw_bf16

• Trainable parameter precision: Pure BF16

• Caption dropout probability: 10.0%

• LoRA Rank: 16

• LoRA Alpha: 16.0

• LoRA Dropout: 0.1

• LoRA initialisation style: default

Datasets

deephouse-512

• Repeats: 10
• Total number of images: 15
• Total number of aspect buckets: 1
• Resolution: 0.262144 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

deephouse-768

• Repeats: 10
• Total number of images: 15
• Total number of aspect buckets: 1
• Resolution: 0.589824 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

deephouse-1024

• Repeats: 10
• Total number of images: 15
• 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

model_id = 'black-forest-labs/FLUX.1-dev'
adapter_id = 'linhqyy/deephouse-st-2911'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "This vibrant photograph captures a woman standing on a wooden deck, gazing out at a stunning, clear, turquoise ocean below. The woman is a tall, slim blonde with long, straight hair flowing down her back. She wears a bright turquoise bikini that contrasts with the vivid blue water and highlights her tanned skin. The wooden deck is sturdy and well-built, with a railing that she leans on for support. The deck is positioned on a cliff that overlooks a breathtaking seascape. The water is a gradient of vibrant turquoise and deep blue, with a sandy beach visible at the bottom of the image. The beach is pristine white, with gentle waves lapping against it. Three sailboats are anchored in the water, adding a sense of scale and perspective. The cliffs are lush with green vegetation, contrasting with the blue water and sky. The sky is a clear, bright blue, with no clouds. The overall mood is one of serene, tranquil relaxation, with the bright colors and clear, sharp details adding to the vividness and clarity of the scene."


## Optional: quantise the model to save on vram.
## Note: The model was not quantised during training, so it is not necessary to quantise it 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=1344,
    height=768,
    guidance_scale=3.0,
).images[0]
image.save("output.png", format="PNG")