Into Non-lieu
Creates surreal non-lieu/non-place style images
Model is trained on images made with Flux Schnell. Dataset's images feature surreal liminal non-lieu (non-place in French) locations. Created images may sometimes give a random Eiffel tower.
Usage guide
Works well with the default settings, also good with the model weight at 0.7 and the Overall prompt weight at 40%. Can be used with other LoRAs and Start images.
Optimal prompting keywords
We recommend using these trigger words when prompting Into Non-lieu to get the best results:
- Surreal still shot
- 80s travel documentary
- liminal non-lieu
No comments yet