Verified: a month ago
SafeTensor
The FLUX.1 [dev] Model is licensed by Black Forest Labs. Inc. under the FLUX.1 [dev] Non-Commercial License. Copyright Black Forest Labs. Inc.
IN NO EVENT SHALL BLACK FOREST LABS, INC. BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.
One of Judy's truest loves is baking, and her skill far surpasses that of generative artificial intelligence. This LoRA brings Flux up to her level by teaching it several basic techniques for scoring sourdough bread. Like scored bread, however, it's a bit rough around the edges, particularly when it comes to side views (will address in V2... soon™).
The primary focus for this LoRA is inpainting, although generation is possible. The full Fill model is slightly more consistent, but the image quality is on par with Q8, so both are viable options. The prompts are visible in the showcase images, and the resources used for each one are listed, so I would encourage you to check them for more guidance.
Finally, be prepared to retry. Some seeds just aren't bready for this much baked goodness yet! I welcome suggestions on how to improve the LoRA - the V1 data set is one image for each caption term, and I am sure that can be added to!
The terms I used in the captions are as follows: Concentric Circles, Curved, Diamond Grid, Grid, Large Cross, Loose Spiral, Single Score, Small Cross, Square, Star, Tight Spiral
As for LoRA strength, it is up to personal taste/what works best for you, but my testing suggests this as a starting point: 0.75 to 1 for curved lines, 1 to 1.5 for straight lines, 1 for both in one image