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: >-
      In the style of a Carravagio painting, Musicians in ornate period clothing
      gathered around a lute player, with scattered sheet music and a violin
      nearby, bathed in dramatic side lighting against a dark background
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_1_0.png
  - text: >-
      In the style of a Carravagio painting, A scholar with a long white beard
      at his desk, quill in hand, with an open book and skull illuminated by a
      single strong light source
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_2_0.png
  - text: >-
      In the style of a Carravagio painting, Three figures playing cards at a
      wooden table, one wearing a feathered hat and striped tunic, another
      peering over shoulders in suspicious observation
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_3_0.png
  - text: >-
      In the style of a Carravagio painting, A figure in white examining their
      reflection in a mirror, surrounded by objects of vanity, with stark
      lighting from the left
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_4_0.png
  - text: >-
      In the style of a Carravagio painting, A street musician with an electric
      guitar performing under a single street lamp, surrounded by watching
      pedestrians with phones creating additional points of light
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_5_0.png
  - text: >-
      In the style of a Carravagio painting, A surgeon in scrubs illuminated by
      surgical lights, bent over their patient, with attending staff in shadows
      around them
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_6_0.png
  - text: >-
      In the style of a Carravagio painting, Poker players at a dimly lit table,
      one in a hoodie checking their phones under the table, while security
      cameras watch from above
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_7_0.png
  - text: >-
      In the style of a Carravagio painting, A food photographer in a dark
      studio, dramatically lit by a single softbox, carefully arranging modern
      cuisine on a rustic wooden table
    parameters:
      negative_prompt: blurry, cropped, ugly
    output:
      url: ./assets/image_8_0.png

Caravaggio-QuarterCrops-Flux-LoRA

This is a standard PEFT LoRA 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: 1408x768
• 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 Carravagio painting, Musicians in ornate period clothing gathered around a lute player, with scattered sheet music and a violin nearby, bathed in dramatic side lighting against a dark background

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of a Carravagio painting, A scholar with a long white beard at his desk, quill in hand, with an open book and skull illuminated by a single strong light source

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of a Carravagio painting, Three figures playing cards at a wooden table, one wearing a feathered hat and striped tunic, another peering over shoulders in suspicious observation

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of a Carravagio painting, A figure in white examining their reflection in a mirror, surrounded by objects of vanity, with stark lighting from the left

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of a Carravagio painting, A street musician with an electric guitar performing under a single street lamp, surrounded by watching pedestrians with phones creating additional points of light

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of a Carravagio painting, A surgeon in scrubs illuminated by surgical lights, bent over their patient, with attending staff in shadows around them

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of a Carravagio painting, Poker players at a dimly lit table, one in a hoodie checking their phones under the table, while security cameras watch from above

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of a Carravagio painting, A food photographer in a dark studio, dramatically lit by a single softbox, carefully arranging modern cuisine on a rustic wooden table

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: 5500

• Learning rate: 8e-05

  • Learning rate schedule: polynomial
  • 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=['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%

• SageAttention: Enabled inference

• LoRA Rank: 64

• LoRA Alpha: None

• LoRA Dropout: 0.1

• LoRA initialisation style: default

Datasets

caravaggio-256

• Repeats: 10
• Total number of images: 45
• Total number of aspect buckets: 5
• Resolution: 0.065536 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

caravaggio-crop-256

• Repeats: 10
• Total number of images: 45
• Total number of aspect buckets: 1
• Resolution: 0.065536 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

caravaggio-512

• Repeats: 10
• Total number of images: 45
• Total number of aspect buckets: 8
• Resolution: 0.262144 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

caravaggio-crop-512

• Repeats: 10
• Total number of images: 45
• Total number of aspect buckets: 1
• Resolution: 0.262144 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

caravaggio-768

• Repeats: 10
• Total number of images: 45
• Total number of aspect buckets: 4
• Resolution: 0.589824 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

caravaggio-crop-768

• Repeats: 10
• Total number of images: 45
• Total number of aspect buckets: 1
• Resolution: 0.589824 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

caravaggio-1024

• Repeats: 10
• Total number of images: 45
• Total number of aspect buckets: 12
• Resolution: 1.048576 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

caravaggio-crop-1024

• Repeats: 10
• Total number of images: 45
• Total number of aspect buckets: 1
• Resolution: 1.048576 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

caravaggio-1440

• Repeats: 10
• Total number of images: 45
• Total number of aspect buckets: 7
• Resolution: 2.0736 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

caravaggio-crop-1440

• Repeats: 10
• Total number of images: 45
• Total number of aspect buckets: 1
• Resolution: 2.0736 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

Inference

import torch
from diffusers import DiffusionPipeline

model_id = 'black-forest-labs/FLUX.1-dev'
adapter_id = 'davidrd123/Caravaggio-QuarterCrops-Flux-LoRA'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

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=1408,
    height=768,
    guidance_scale=3.0,
).images[0]
image.save("output.png", format="PNG")