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.
This LoRA collection works better for INPAINT, but it does a good job on txt2img too.
I was having a problem generating images with people who have piercings (nose or lips). Flux doesn't know how to handle that properly, and even training a LoRA for a person with piercings resulted in awful results.
I found some piercing LoRAs, but I noticed a common problem: Flux needs to be trained on the specific side of the face where the piercing should be, otherwise Flux will put the piercing on both sides.
So, I trained it focusing on the correct sides, trying to fix this. It's not perfect, but it works well!
Trained so far - SELECT THE TYPES AT THE PAGE TOP:
ring nose piercing right side
ring nose piercing left side
pin nose piercing right side
pin nose piercing left side
ring nose septum u-shape
lip piercing centered
eyebrows piercing right side
eyebrows piercing left side
medusa piercing centered
lip piercing right side
lip piercing left side
Work in progress:
double lip piercing right side
double lip piercing left side
Because the training data used close-up shots (to avoid changing generated the faces), it might not work as well when using txt2img with faces that aren't in close-up.
In that case, please use inpainting, configuring the bounding box as a face close-up. The process to do that may vary with other tools; I'm using InvokeAI.
