playerzer0x/002-dsyjssbrhr_20250212_013300_1e-4_100_bs8_ao-adamw8bit

002-dsyjssbrhr202502120133001e-4100bs8ao-adamw8bit

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

The main validation prompt used during training was:

dsyjssbrhr style

Validation settings

• —CFG: 3.5
• —CFG Rescale: 0.0
• —Steps: 28
• —Sampler: FlowMatchEulerDiscreteScheduler
• —Seed: 69
• —Resolution: 1024x1024
• —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:

<Gallery />

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

Training settings

• —Training epochs: 8
• —Training steps: 100
• —Learning rate: 0.0002
• —Learning rate schedule: constant
• —Warmup steps: 0
• —Max grad norm: 2.0
• —Effective batch size: 8
• —Micro-batch size: 2
• —Gradient accumulation steps: 1
• —Number of GPUs: 4
• —Gradient checkpointing: True
• —Prediction type: flow-matching (extra parameters=['shift=3', 'fluxguidancemode=constant', 'fluxguidancevalue=1.0', 'flowmatchingloss=compatible', 'fluxloratarget=all+ffs'])
• —Optimizer: ao-adamw8bitweight_decay=1e-3
• —Trainable parameter precision: Pure BF16
• —Caption dropout probability: 5.0%

• —LoRA Rank: 16
• —LoRA Alpha: 16.0
• —LoRA Dropout: 0.1
• —LoRA initialisation style: default

Datasets

002-dsyjssbrhr-512

• —Repeats: 0
• —Total number of images: ~28
• —Total number of aspect buckets: 3
• —Resolution: 0.262144 megapixels
• —Cropped: False
• —Crop style: None
• —Crop aspect: None
• —Used for regularisation data: No

002-dsyjssbrhr-768

• —Repeats: 0
• —Total number of images: ~24
• —Total number of aspect buckets: 1
• —Resolution: 0.589824 megapixels
• —Cropped: False
• —Crop style: None
• —Crop aspect: None
• —Used for regularisation data: No

002-dsyjssbrhr-1024

• —Repeats: 0
• —Total number of images: ~28
• —Total number of aspect buckets: 2
• —Resolution: 1.048576 megapixels
• —Cropped: False
• —Crop style: None
• —Crop aspect: None
• —Used for regularisation data: No

Inference

import torch
from diffusers import DiffusionPipeline

model_id = 'black-forest-labs/FLUX.1-dev'
adapter_id = 'playerzer0x/002-dsyjssbrhr_20250212_013300_1e-4_100_bs8_ao-adamw8bit'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "dsyjssbrhr style"


## Optional: quantise the model to save on vram.
## Note: The model was not quantised during training, so it is not necessary to quantise it 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=28,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(69),
    width=1024,
    height=1024,
    guidance_scale=3.5,
).images[0]
image.save("output.png", format="PNG")