Text-to-Image
Diffusers
sd3
sd3-diffusers
simpletuner
Not-For-All-Audiences
lora
template:sd-lora
lycoris
metadata
license: other
base_model: stabilityai/stable-diffusion-3.5-large
tags:
- sd3
- sd3-diffusers
- text-to-image
- diffusers
- simpletuner
- not-for-all-audiences
- 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: >-
A fusion of Furkan and Megamind, keeping Furkan’s curly hair, glasses, and
beard but giving him blue skin and a large head. He wears a sleek black
leather suit with a cape, exuding confidence and intellect. Standing in a
neon-lit futuristic city, he’s ready to conquer the world with both charm
and genius.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_1_0.png
- text: >-
Dr. Furkan, a brilliant AI researcher with many PhDs and papers. He wears
a lab coat, filled with pens and gadgets, and keeps his signature curly
hair, beard, and glasses. In a futuristic lab surrounded by holograms and
robots, he’s deep in thought, crafting the next breakthrough in AI.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_2_0.png
- text: >-
Furkan as an eccentric pizza chef, tossing glowing dough into the air with
flair. He wears a flour-dusted apron and a tall chef’s hat, grinning as if
about to reveal a bizarre new recipe. The kitchen is chaotic, with flying
ingredients and robots helping him prepare his masterpiece.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_3_0.png
- text: >-
Furkan as a wild rockstar, dressed in leather with a guitar slung over his
shoulder. His curly hair spills from under a spiked crown, and his beard
has glowing streaks. On stage, bathed in neon lights, he shreds a guitar
solo while the crowd roars, glasses reflecting the spotlight.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_4_0.png
- text: >-
furkan wearing a large turban wrapped carefully in multiple layers,
creating a tall and imposing shape, the turban is adorned with intricate
patterns reflecting the traditional art of the region, his robe is
embellished with detailed embroidery especially on the chest and along the
sleeves, his cloak appears to be embroidered as well with geometric
designs, his gaze is direct and determined conveying a sense of pride and
authority
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_5_0.png
- text: >-
Furkan standing atop a windswept dune, his cloak billowing behind him in
the desert breeze, the sun casts a golden glow highlighting the intricate
beadwork on his vest, his turban is wrapped elegantly with a feather
tucked into the folds, his boots are sturdy yet ornate with silver
buckles, his gaze is fixed on the horizon embodying a spirit of adventure
and resilience.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_6_0.png
- text: >-
Furkan seated beside a tranquil river under the shade of ancient willow
trees, his lute rests gently on his lap adorned with mother-of-pearl
inlays, his tunic is woven with threads of gold forming delicate floral
patterns, his turban is modest yet dignified signifying wisdom, his
fingers poised to strum a melody, his eyes reflect a deep serenity and
connection with nature.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_7_0.png
- text: >-
Furkan walking through a bustling bazaar filled with vibrant colors and
enticing aromas, his robes are layered and flowing embellished with
patterns of swirling calligraphy, his turban is crowned with a precious
gem catching the light, he carries a staff carved with symbols of
protection, his smile is warm and welcoming exuding generosity and
leadership.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_8_0.png
- text: >-
Furkan riding a majestic horse across verdant plains under a clear sky,
his armor is crafted with meticulous detail showcasing craftsmanship of
the highest order, his turban is secured tightly ready for the journey
ahead, his cape features emblems of his lineage, his posture is upright
and commanding illustrating nobility and courage.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_9_0.png
- text: >-
Furkan meditating atop a sacred mountain enveloped in mist, his robes are
simple yet adorned with sacred texts along the hems, his turban is pure
white symbolizing enlightenment, around his neck hangs a pendant inscribed
with ancient runes, his hands rest gently on his knees, his expression is
peaceful and introspective indicating spiritual depth and wisdom.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_10_0.png
- text: >-
A young woman dancing gracefully in a meadow blanketed with wildflowers,
her dress flows effortlessly adorned with lace and delicate embroidery,
her hair is woven with blossoms reflecting the natural beauty around her,
she wears a pendant shaped like a crescent moon, her movements are fluid
and expressive conveying joy and freedom.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_11_0.png
- text: >-
An elderly man crafting a wooden sculpture in his rustic workshop filled
with tools and wood shavings, his apron is worn and stained telling
stories of years dedicated to his craft, his hands are steady and skilled
carving intricate details, sunlight streams through the window
illuminating dust particles, his eyes are focused and content embodying
dedication and mastery.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_12_0.png
- text: >-
A child exploring a secret garden hidden behind an old gate covered in
ivy, wearing overalls with patches in various shapes and colors, carrying
a small satchel filled with collected treasures like stones and feathers,
their shoes are muddy from adventure, their face lights up with curiosity
and wonder capturing the essence of innocence and discovery.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_13_0.png
- text: >-
A scholar engrossed in ancient texts within a grand library lined with
towering bookshelves, their spectacles rest low on their nose, a quill in
hand poised over parchment filled with notes, their robe is simple but
accented with a finely woven scarf, candles flicker casting a warm glow,
their expression is thoughtful and analytical demonstrating intellect and
passion for knowledge.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_14_0.png
- text: >-
An artist painting a vibrant mural on a city wall, wearing a jumpsuit
splattered with a spectrum of colors, their hair is tied back revealing
earrings shaped like tiny palettes, a collection of brushes protrudes from
their pocket, passersby stop to admire the emerging masterpiece, their
face shows determination and creativity highlighting the transformative
power of art.
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_15_0.png
- text: A photo-realistic image of a cat
parameters:
negative_prompt: blurry, cropped, ugly
output:
url: ./assets/image_16_0.png
Furkan-SD3
This is a LyCORIS adapter derived from stabilityai/stable-diffusion-3.5-large.
The main validation prompt used during training was:
A photo-realistic image of a cat
Validation settings
- CFG:
4.0
- CFG Rescale:
0.0
- Steps:
20
- Sampler:
None
- Seed:
42
- Resolution:
1024x1024
Note: The validation settings are not necessarily the same as the training settings.
You can find some example images in the following gallery:
The text encoder was not trained. You may reuse the base model text encoder for inference.
Training settings
- Training epochs: 2
- Training steps: 2400
- Learning rate: 1e-05
- Max grad norm: 0.01
- Effective batch size: 3
- Micro-batch size: 1
- Gradient accumulation steps: 1
- Number of GPUs: 3
- Prediction type: flow-matching
- Rescaled betas zero SNR: False
- Optimizer: optimi-lion
- Precision: Pure BF16
- Quantised: No
- Xformers: Not used
- LyCORIS Config:
{
"bypass_mode": true,
"algo": "lokr",
"multiplier": 1.0,
"full_matrix": true,
"linear_dim": 10000,
"linear_alpha": 1,
"factor": 12,
"apply_preset": {
"target_module": [
"Attention"
],
"module_algo_map": {
"Attention": {
"factor": 6
}
}
}
}
Datasets
reg-512
- Repeats: 0
- Total number of images: ~288
- Total number of aspect buckets: 11
- Resolution: 0.262144 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: Yes
reg-1024
- Repeats: 0
- Total number of images: ~291
- Total number of aspect buckets: 14
- Resolution: 1.048576 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: Yes
furkan-uncut-512
- Repeats: 30
- Total number of images: ~18
- Total number of aspect buckets: 1
- Resolution: 0.262144 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No
furkan-cropped-512
- Repeats: 30
- Total number of images: ~18
- Total number of aspect buckets: 3
- Resolution: 0.262144 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No
furkan-uncut-1024
- Repeats: 30
- Total number of images: ~12
- Total number of aspect buckets: 1
- Resolution: 1.048576 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No
furkan-cropped-1024
- Repeats: 30
- Total number of images: ~12
- 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
from lycoris import create_lycoris_from_weights
model_id = 'stabilityai/stable-diffusion-3.5-large'
adapter_id = 'pytorch_lora_weights.safetensors' # you will have to download this manually
lora_scale = 1.0
wrapper, _ = create_lycoris_from_weights(lora_scale, adapter_id, pipeline.transformer)
wrapper.merge_to()
prompt = "A photo-realistic image of a cat"
negative_prompt = 'blurry, cropped, ugly'
pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu')
image = pipeline(
prompt=prompt,
negative_prompt=negative_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=4.0,
).images[0]
image.save("output.png", format="PNG")