Trigger: s3l3n3

This is a quick test to see how quick I can train on my 4060Ti 16Gb.
Quality is not super great, but generally better than my SDXL tests.

Trained for 60 epochs, 1 repeat for a total of 600 steps, 512x512, Kohya Gui SD3-Flux branch using flux preset and adjusting to network/alpha of 4. About 32mins for 10 low quality pictures of different resolutions and aspect ratios.

Better likeness can be achieved by describing her look in the film, some words to help: Black hair, black leather suit.

Really amazed by Flux's capabilities.

Trigger words

You should use s3l3n3 to trigger the image generation.

Download model

Weights for this model are available in Safetensors format.

Download them in the Files & versions tab.

Use it with the 🧨 diffusers library

from diffusers import AutoPipelineForText2Image
import torch

device = "cuda" if torch.cuda.is_available() else "cpu"

pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.bfloat16).to(device)
pipeline.load_lora_weights('Keltezaa/selene-underworld-kate-beckinsale-flux-lora-test', weight_name='Selene Underworld (s3l3n3) - Flux test-000060.safetensors')
image = pipeline('`s3l3n3`').images[0]

For more details, including weighting, merging and fusing LoRAs, check the documentation on loading LoRAs in diffusers