Ben-Brand-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: 1024x1024
• Skip-layer guidance:

Note: The validation settings are not necessarily the same as the training settings.

Prompt

In the style of a b3nbr4nd painting, Massive green serpent, coiled around a cracked stone obelisk, golden eyes glowing, dark blue scales with iridescent shimmer, open mouth revealing sharp fangs, pink sky fading into deep purple, sand dunes stretching into the horizon, ancient ruins partially buried, red banners flapping in the wind.

Prompt

In the style of a b3nbr4nd painting, The Fractured Cathedral – Ruined temple standing between timelines, stained glass windows refracting multiple realities, golden gears turning in the vaulted ceiling, priests in robes of shifting colors, a mechanical choir humming in binary, relics of forgotten AI scattered on an altar, static crackling like divine whispers.

Prompt

In the style of a b3nbr4nd painting, The Cartographer of Lost Time – A hunched figure tracing glowing lines across an ancient map, ink shifting as if alive, continents forming and vanishing, thousands of tiny golden orbs orbiting the parchment, the map itself whispering of places that no longer exist, candlelight flickering in unknown patterns.

Prompt

In the style of a b3nbr4nd painting, A steaming outdoor pool carved from volcanic rock, floating lanterns casting rippling golden reflections, pale steam curling upwards into a canopy of sapphire sky, koi fish with silver scales swimming in slow, deliberate circles.

Prompt

In the style of a b3nbr4nd painting, snake, black snakes, desert, cacti, cactus, prickly pear cactus, yucca plant, orange eye, mountains, purple mountains, blue sky, green cactus, yellow desert, close-up view, coiled snake, landscape, arid environment, vegetation, plant, side view, foreground snake, background mountains

Prompt

In the style of a b3nbr4nd painting, starry night sky, tall buildings, skyscrapers, windows, connected buildings, walkway, green checkerboard pattern walkway, arched windows, trees with lights, urban landscape, geometric architecture, city night scene, multicolored buildings, modern cityscape, midground walkway, background buildings, illuminated windows, high-rise structure, blue building, city park with lights

Prompt

In the style of a b3nbr4nd painting, Massive gold-plated pyramid, intricate carvings on its surface, glowing blue symbols along edges, deep violet sky, swirling pink clouds, hovering metallic spheres reflecting the landscape, a lone traveler in red robes ascending marble stairs, crystal-clear river flowing around pyramid base.

Prompt

In the style of a b3nbr4nd painting, blue vases, pink tulip, table with patterned tablecloth, orange mugs, green teapot, posters on wall, red poster, text on red poster, white text, yellow wall, drawers in background, multiple cups, ceramic kettle, lid on kettle, floral mug, window with striped curtains, indoor scene, objects on table, vase in center, tea set

Prompt

In the style of a b3nbr4nd painting, Armored knight in reflective silver plate, tattered blue cape, standing on a vast frozen lake, sword planted in ice, glowing runes along blade, towering ice formations in background, pale full moon overhead, scattered red leaves on ice surface, distant torches flickering along mountain ridge.

Prompt

In the style of a b3nbr4nd painting, Towering white marble golem, deep blue glowing veins across its body, massive stone hands gripping a golden spear, standing before an ancient archway, intricate carvings on stone, overgrown vines wrapping around pillars, golden sunset casting long shadows, wind carrying red and orange leaves through the air.

Prompt

In the style of a b3nbr4nd painting, Giant bird made of sapphire flames, wings spread wide, glowing embers trailing behind, deep black sky with thousands of stars, large yellow moon behind bird, floating islands in the distance, waterfalls spilling into nothingness, golden temple on one of the islands, reflections of stars on the bird’s feathers.

The text encoder was not trained. You may reuse the base model text encoder for inference.

Training settings

• Training epochs: 2

• Training steps: 3750

• Learning rate: 0.00015

  • Learning rate schedule: constant
  • Warmup steps: 100

• Max grad norm: 0.1

• Effective batch size: 6

  • Micro-batch size: 2
  • Gradient accumulation steps: 3
  • 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: 64

• LoRA Alpha: None

• LoRA Dropout: 0.1

• LoRA initialisation style: default

Datasets

ben-brand-256

• Repeats: 10
• Total number of images: 98
• Total number of aspect buckets: 3
• Resolution: 0.065536 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

ben-brand-crop-256

• Repeats: 10
• Total number of images: 98
• Total number of aspect buckets: 1
• Resolution: 0.065536 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

ben-brand-512

• Repeats: 10
• Total number of images: 98
• Total number of aspect buckets: 3
• Resolution: 0.262144 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

ben-brand-crop-512

• Repeats: 10
• Total number of images: 98
• Total number of aspect buckets: 1
• Resolution: 0.262144 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

ben-brand-768

• Repeats: 10
• Total number of images: 98
• Total number of aspect buckets: 3
• Resolution: 0.589824 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

ben-brand-crop-768

• Repeats: 10
• Total number of images: 98
• Total number of aspect buckets: 1
• Resolution: 0.589824 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

ben-brand-1024

• Repeats: 10
• Total number of images: 98
• Total number of aspect buckets: 4
• Resolution: 1.048576 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

ben-brand-crop-1024

• Repeats: 10
• Total number of images: 98
• Total number of aspect buckets: 1
• Resolution: 1.048576 megapixels
• Cropped: True
• Crop style: center
• Crop aspect: square
• Used for regularisation data: No

ben-brand-1440

• Repeats: 10
• Total number of images: 98
• Total number of aspect buckets: 2
• Resolution: 2.0736 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None
• Used for regularisation data: No

ben-brand-crop-1440

• Repeats: 10
• Total number of images: 98
• 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/Ben-Brand-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=1024,
    height=1024,
    guidance_scale=3.0,
).images[0]
image.save("output.png", format="PNG")