HarakaBaraka/meg-flux-lora

Meg the dog - Flux LoRA

This is a LoRA (Low-Rank Adaptation) model trained on images of Meg the dog using the Flux.1-dev base model.

Usage

from diffusers import DiffusionPipeline
import torch

# Load the base model
pipe = DiffusionPipeline.from_pretrained(
    "black-forest-labs/FLUX.1-dev",
    torch_dtype=torch.bfloat16
).to("cuda")

# Load the LoRA weights
pipe.load_lora_weights("HarakaBaraka/meg-flux-lora")

# Generate an image
prompt = "a photo of Meg the dog"
image = pipe(prompt, num_inference_steps=50, guidance_scale=6).images[0]
image.save("generated_image.png")

Training Details

• Base model: black-forest-labs/FLUX.1-dev
• Training steps: 500
• Learning rate: 0.0004
• LoRA rank: 16
• Resolution: 512px

Example Prompts

Try these prompts with your model:

• Meg the dog
• a painting of Meg the dog in the style of Van Gogh
• Meg the dog wearing a superhero costume
• oil painting of Meg the dog flying through space as an astronaut

Training Data

This model was trained on personal images. Please use responsibly and respect privacy.