copycat-flux-lora-training

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

The main validation prompt used during training was:

NSFW, Realistic anime style, Young woman with long red hair and flower hairpin standing on balcony at sunset, looking at viewer with blushing cheeks and wide blue eyes. White ruffled crop top exposing midriff and a white high-waist skirt that that exposes her pussy. No panties, exposed pussy visible. Visible navel. Bottomless with pale thighs and smooth skin. City skyline in background with church spire. Upper body view.

Validation settings

CFG: 3.5
CFG Rescale: 0.0
Steps: 15
Sampler: None
Seed: 42
Resolution: 768x1368

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
'

Prompt
NSFW, Realistic anime style, Young woman with long red hair and flower hairpin standing on balcony at sunset, looking at viewer with blushing cheeks and wide blue eyes. White ruffled crop top exposing midriff and a white high-waist skirt that that exposes her pussy. No panties, exposed pussy visible. Visible navel. Bottomless with pale thighs and smooth skin. City skyline in background with church spire. Upper body view.

Negative Prompt
'

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

Training settings

Training epochs: 39
Training steps: 1000
Learning rate: 0.0001
Effective batch size: 4
- Micro-batch size: 4
- Gradient accumulation steps: 1
- Number of GPUs: 1
Prediction type: flow-matching
Rescaled betas zero SNR: False
Optimizer: adamw_bf16
Precision: Pure BF16
Quantised: No
Xformers: Not used
LoRA Rank: 64
LoRA Alpha: None
LoRA Dropout: 0.1
LoRA initialisation style: default

Datasets

default_dataset_arb

Repeats: 0
Total number of images: 97
Total number of aspect buckets: 1
Resolution: 1.0 megapixels
Cropped: False
Crop style: None
Crop aspect: None

Inference

import torch
from diffusers import DiffusionPipeline

model_id = 'black-forest-labs/FLUX.1-dev'
adapter_id = 'BeetleBonsai798/copycat-flux-lora-training'
pipeline = DiffusionPipeline.from_pretrained(model_id)
pipeline.load_lora_weights(adapter_id)

prompt = "NSFW, Realistic anime style, Young woman with long red hair and flower hairpin standing on balcony at sunset, looking at viewer with blushing cheeks and wide blue eyes. White ruffled crop top exposing midriff and a white high-waist skirt that that exposes her pussy. No panties, exposed pussy visible. Visible navel. Bottomless with pale thighs and smooth skin. City skyline in background with church spire. Upper body view."

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=768,
    height=1368,
    guidance_scale=3.5,
).images[0]
image.save("output.png", format="PNG")