lora-Maurice-Prendergast-Flux

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

The main validation prompt used during training was:

mp_style, Street scene, 50 figures (many women: colorful dresses, many men: suits), 23 umbrellas (orange, red, yellow, green), bridge, buildings background, water, boats, Italian flag, steps, lamps, crowd ascending descending bridge, signature bottom-left

Validation settings

• CFG: 7.5
• CFG Rescale: 0.0
• Steps: 20
• Sampler: None
• Seed: 42
• Resolution: 1024

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

mp_style, Street scene, 50 figures (many women: colorful dresses, many men: suits), 23 umbrellas (orange, red, yellow, green), bridge, buildings background, water, boats, Italian flag, steps, lamps, crowd ascending descending bridge, signature bottom-left

Negative Prompt

blurry, cropped, ugly

Prompt

mp_style, Outdoor setting, 30 figures, mostly women in long dresses, some men, some children, 14 umbrellas (orange, red, blue), large Gothic arches, upper facade with decorative patterns, stone pavement, black lamp post, signature bottom-left.

Negative Prompt

blurry, cropped, ugly

Prompt

mp_style, park scene, fountain, autumn foliage, orange, yellow, red, trees, people, white chairs, standing, blue sky, clouds, vibrant colors, peaceful, serene.

Negative Prompt

blurry, cropped, ugly

Prompt

mp_style, Street scene, 50 figures (many women: colorful dresses, many men: suits), 23 umbrellas (orange, red, yellow, green), bridge, buildings background, water, boats, Italian flag, steps, lamps, crowd ascending descending bridge, signature bottom-left

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: 235
• Training steps: 4000
• Learning rate: 0.0004
• Effective batch size: 6
  • Micro-batch size: 6
  • Gradient accumulation steps: 1
  • Number of GPUs: 1
• Prediction type: flow-matching
• Rescaled betas zero SNR: False
• Optimizer: AdamW, stochastic bf16
• Precision: Pure BF16
• Xformers: Not used
• LoRA Rank: 64
• LoRA Alpha: None
• LoRA Dropout: 0.1
• LoRA initialisation style: default

Datasets

MauricePrendergastRedo3

• Repeats: 0
• Total number of images: 102
• Total number of aspect buckets: 1
• Resolution: 1024 px
• Cropped: True
• Crop style: center
• Crop aspect: square

Inference

import torch
from diffusers import DiffusionPipeline

model_id = 'black-forest-labs/FLUX.1-dev'
adapter_id = 'davidrd123/lora-Maurice-Prendergast-Flux'
pipeline = DiffusionPipeline.from_pretrained(model_id)
pipeline.load_lora_weights(adapter_id)

prompt = "mp_style, Street scene, 50 figures (many women: colorful dresses, many men: suits), 23 umbrellas (orange, red, yellow, green), bridge, buildings background, water, boats, Italian flag, steps, lamps, crowd ascending descending bridge, signature bottom-left"


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=7.5,
).images[0]
image.save("output.png", format="PNG")