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: >-
DRK painting, A woman in traditional dress with a long skirt and apron
stands beside a peaceful river, holding a wooden staff. She wears a white
blouse and brown bodice, with a red headscarf. Lush wildflowers bloom
nearby, and distant hills roll under a cloudy sky
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_1_0.png
- text: >-
DRK painting, A woman kneels in a flourishing garden filled with colorful
flowers, wearing a striped skirt and white blouse with an apron. She holds
gardening tools and a basket of freshly picked flowers rests beside her. A
small boat is visible on the river in the background
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_2_0.png
- text: >-
DRK painting, Two women walk along a country path near stone walls, one
carrying a large water jug and the other holding a basket of vegetables.
They wear traditional dresses with aprons and headscarves. Bare trees and
distant cottages complete the rural scene
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_3_0.png
- text: >-
DRK painting, A young woman in a casual sundress tends to an urban rooftop
garden, surrounded by potted plants and hanging vines. She holds a modern
watering can while looking out over the city skyline at sunset
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_4_0.png
- text: >-
DRK painting, A solitary figure in flowing clothes stands on rocky coastal
cliffs, holding a telescope and gazing out at stormy seas. Their garments
blow in the wind as seabirds wheel overhead and waves crash below
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_5_0.png
- text: >-
DRK painting, A street vendor arranges colorful flowers at their modern
market stall, wearing a practical apron over contemporary clothes. Glass
buildings reflect the morning light behind them while early customers
browse the vibrant blooms
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_6_0.png
Flux-Daniel-Ridgway-Knight-LoKr
This is a LyCORIS adapter 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:
896x1152 - 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
- DRK painting, A woman in traditional dress with a long skirt and apron stands beside a peaceful river, holding a wooden staff. She wears a white blouse and brown bodice, with a red headscarf. Lush wildflowers bloom nearby, and distant hills roll under a cloudy sky
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- DRK painting, A woman kneels in a flourishing garden filled with colorful flowers, wearing a striped skirt and white blouse with an apron. She holds gardening tools and a basket of freshly picked flowers rests beside her. A small boat is visible on the river in the background
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- DRK painting, Two women walk along a country path near stone walls, one carrying a large water jug and the other holding a basket of vegetables. They wear traditional dresses with aprons and headscarves. Bare trees and distant cottages complete the rural scene
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- DRK painting, A young woman in a casual sundress tends to an urban rooftop garden, surrounded by potted plants and hanging vines. She holds a modern watering can while looking out over the city skyline at sunset
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- DRK painting, A solitary figure in flowing clothes stands on rocky coastal cliffs, holding a telescope and gazing out at stormy seas. Their garments blow in the wind as seabirds wheel overhead and waves crash below
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- DRK painting, A street vendor arranges colorful flowers at their modern market stall, wearing a practical apron over contemporary clothes. Glass buildings reflect the morning light behind them while early customers browse the vibrant blooms
- 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: 10
Training steps: 14000
Learning rate: 8e-05
- Learning rate schedule: constant
- 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'])
Optimizer: adamw_bf16
Trainable parameter precision: Pure BF16
Caption dropout probability: 10.0%
SageAttention: Enabled inference
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
drk-256
- Repeats: 10
- Total number of images: 32
- Total number of aspect buckets: 1
- Resolution: 0.065536 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No
drk-crop-256
- Repeats: 10
- Total number of images: 32
- Total number of aspect buckets: 1
- Resolution: 0.065536 megapixels
- Cropped: True
- Crop style: center
- Crop aspect: square
- Used for regularisation data: No
drk-512
- Repeats: 10
- Total number of images: 32
- Total number of aspect buckets: 3
- Resolution: 0.262144 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No
drk-crop-512
- Repeats: 10
- Total number of images: 32
- Total number of aspect buckets: 1
- Resolution: 0.262144 megapixels
- Cropped: True
- Crop style: center
- Crop aspect: square
- Used for regularisation data: No
drk-768
- Repeats: 10
- Total number of images: 32
- Total number of aspect buckets: 3
- Resolution: 0.589824 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No
drk-crop-768
- Repeats: 10
- Total number of images: 32
- Total number of aspect buckets: 1
- Resolution: 0.589824 megapixels
- Cropped: True
- Crop style: center
- Crop aspect: square
- Used for regularisation data: No
drk-1024
- Repeats: 10
- Total number of images: 32
- Total number of aspect buckets: 9
- Resolution: 1.048576 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No
drk-crop-1024
- Repeats: 10
- Total number of images: 32
- Total number of aspect buckets: 1
- Resolution: 1.048576 megapixels
- Cropped: True
- Crop style: center
- Crop aspect: square
- Used for regularisation data: No
drk-1440
- Repeats: 10
- Total number of images: 32
- Total number of aspect buckets: 10
- Resolution: 2.0736 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No
drk-crop-1440
- Repeats: 10
- Total number of images: 32
- 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
from lycoris import create_lycoris_from_weights
def download_adapter(repo_id: str):
import os
from huggingface_hub import hf_hub_download
adapter_filename = "pytorch_lora_weights.safetensors"
cache_dir = os.environ.get('HF_PATH', os.path.expanduser('~/.cache/huggingface/hub/models'))
cleaned_adapter_path = repo_id.replace("/", "_").replace("\\", "_").replace(":", "_")
path_to_adapter = os.path.join(cache_dir, cleaned_adapter_path)
path_to_adapter_file = os.path.join(path_to_adapter, adapter_filename)
os.makedirs(path_to_adapter, exist_ok=True)
hf_hub_download(
repo_id=repo_id, filename=adapter_filename, local_dir=path_to_adapter
)
return path_to_adapter_file
model_id = 'black-forest-labs/FLUX.1-dev'
adapter_repo_id = 'davidrd123/Flux-Daniel-Ridgway-Knight-LoKr'
adapter_filename = 'pytorch_lora_weights.safetensors'
adapter_file_path = download_adapter(repo_id=adapter_repo_id)
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
lora_scale = 1.0
wrapper, _ = create_lycoris_from_weights(lora_scale, adapter_file_path, pipeline.transformer)
wrapper.merge_to()
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=896,
height=1152,
guidance_scale=3.0,
).images[0]
image.save("output.png", format="PNG")