Trigger is "Weirdcore aesthetic." at the start of the prompt. For the first sample image, I had to manually fix the text using Photopea and inpainting with a low denoise value.

I'm not very experienced with making natural language datasets, so if any of you think you could improve on this, the dataset is available too.

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)
pipe.load_lora_weights('analogspiderweb/weirdcore-aesthetic-flux-dev', weight_name='weirdcore-aesthetic-flux-dev.safetensors')
image = pipeline('Weirdcore aesthetic. Low quality photo of an empty room with a giant hole in the floor. The hole leads to a beautiful field with edited butterflies and rainbows coming out. The blue text says "don\'t you wanna see what\'s on the other side?". ').images[0]

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