flux-training-losercity-next
This is a LoRA derived from black-forest-labs/FLUX.1-dev.
The main validation prompt used during training was:
loona from helluva boss is eating a donut
Validation settings
- CFG:
3.5 - CFG Rescale:
0.0 - Steps:
15 - 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
- ddttfm, A steaming bowl of ramen with chopsticks resting on the edge.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A child holding a colorful kite in a park on a sunny day.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, An artist painting a canvas with bright, abstract colors in a studio.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A stack of old vinyl records next to a vintage turntable.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A golden retriever playing fetch with a red ball in the backyard.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A woman in a red dress dancing gracefully under a spotlight.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A man in a suit typing on a laptop at a modern office desk.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A cup of tea with a lemon slice on a wooden table with a book beside it.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A bicycle parked against a brick wall covered in ivy.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A chef garnishing a gourmet dish with fresh herbs in a restaurant kitchen.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A cozy reading nook with a green armchair and a stack of novels.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A blue suitcase with travel stickers and a passport peeking out.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A street artist painting a vibrant mural on a city wall.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A dog wearing a superhero cape playing in the backyard.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A set of art supplies including brushes, paints, and a sketchbook on a table.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A man in casual clothes relaxing with a fishing rod by a serene lake.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A woman in a yoga pose on a beach at sunrise.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A futuristic city skyline with tall, illuminated skyscrapers at dusk.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A vintage typewriter with a blank sheet of paper ready to be typed on.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, A collection of colorful balloons floating in the sky at a party.
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, Small green frog napping on the nose of a golden retriever
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- ddttfm, Small tree frog perching on the tip of a golden retriever's nose
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- loona from helluva boss is eating a donut
- 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: 466
- Training steps: 2800
- Learning rate: 0.0001
- 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: Enabled
- LoRA Rank: 64
- LoRA Alpha: None
- LoRA Dropout: 0.1
- LoRA initialisation style: default
Datasets
smart_tuner
- Repeats: 0
- Total number of images: 36
- Total number of aspect buckets: 1
- Resolution: 1 megapixels
- 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 = 'Dnau15/flux-training-losercity-next'
pipeline = DiffusionPipeline.from_pretrained(model_id)
pipeline.load_lora_weights(adapter_id)
prompt = "loona from helluva boss is eating a donut"
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=15,
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=3.5,
).images[0]
image.save("output.png", format="PNG")
- Downloads last month
- 20
Model tree for Dnau15/flux-training-losercity-next
Base model
black-forest-labs/FLUX.1-dev