train-realistic-16-09-v2

train-realistic-16-09-v2

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

The main validation prompt used during training was:

This idyllic alpine scene captures a serene lake nestled between lush green hills and towering mountains capped with snow. A wooden chalet, located on a gently sloping hill, overlooks the vibrant turquoise waters of the lake, its spacious balcony offering sweeping views of the breathtaking landscape. The surrounding meadows are dotted with colorful wildflowers, and tall trees line the hillsides, blending harmoniously with the forested mountains in the distance. Beyond the lake, a small village can be seen, framed by majestic peaks that rise sharply against a bright blue sky adorned with fluffy white clouds. The crystal-clear water reflects the sky and the surrounding greenery, enhancing the tranquility of the scene. Every detail, from the texture of the grass to the snow on the distant mountains, is rendered with vibrant clarity, evoking a sense of peace and natural beauty in this picturesque mountain retreat.

Validation settings

• CFG: 4.0
• CFG Rescale: 0.0
• Steps: 20
• Sampler: None
• Seed: 42
• Resolution: 1344x768

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: 14
• Training steps: 9000
• Learning rate: 8e-05
• Effective batch size: 1
  • Micro-batch size: 1
  • Gradient accumulation steps: 1
  • Number of GPUs: 1
• Prediction type: flow-matching
• Rescaled betas zero SNR: False
• Optimizer: adamw_bf16
• Precision: Pure BF16
• Quantised: Yes: int8-quanto
• Xformers: Not used
• LoRA Rank: 64
• LoRA Alpha: None
• LoRA Dropout: 0.1
• LoRA initialisation style: default

Datasets

realistic-16-09-512

• Repeats: 10
• Total number of images: 14
• Total number of aspect buckets: 1
• Resolution: 0.262144 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None

realistic-16-09-1024

• Repeats: 10
• Total number of images: 14
• Total number of aspect buckets: 1
• Resolution: 1.048576 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None

realistic-16-09-512-crop

• Repeats: 10
• Total number of images: 14
• Total number of aspect buckets: 1
• Resolution: 0.262144 megapixels
• Cropped: True
• Crop style: random
• Crop aspect: square

realistic-16-09-1024-crop

• Repeats: 10
• Total number of images: 14
• Total number of aspect buckets: 1
• Resolution: 1.048576 megapixels
• Cropped: True
• Crop style: random
• Crop aspect: square

Inference

import torch
from diffusers import DiffusionPipeline

model_id = 'black-forest-labs/FLUX.1-dev'
adapter_id = 'datnt114/train-realistic-16-09-v2'
pipeline = DiffusionPipeline.from_pretrained(model_id)
pipeline.load_lora_weights(adapter_id)

prompt = "This idyllic alpine scene captures a serene lake nestled between lush green hills and towering mountains capped with snow. A wooden chalet, located on a gently sloping hill, overlooks the vibrant turquoise waters of the lake, its spacious balcony offering sweeping views of the breathtaking landscape. The surrounding meadows are dotted with colorful wildflowers, and tall trees line the hillsides, blending harmoniously with the forested mountains in the distance. Beyond the lake, a small village can be seen, framed by majestic peaks that rise sharply against a bright blue sky adorned with fluffy white clouds. The crystal-clear water reflects the sky and the surrounding greenery, enhancing the tranquility of the scene. Every detail, from the texture of the grass to the snow on the distant mountains, is rendered with vibrant clarity, evoking a sense of peace and natural beauty in this picturesque mountain retreat."

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=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=1344,
    height=768,
    guidance_scale=4.0,
).images[0]
image.save("output.png", format="PNG")