Sign In

Jellyfish Hair - Danrom

9

Download

1 variant available

SafeTensor

124.56 MB

Verified:

Type

LoRA

Stats

97

9

32

Reviews

Published

Apr 6, 2026

Base Model

Flux.2 Klein 9B

Hash

AutoV2
E6209D3F79

Trigger Words

jfhaircut
default creator card background decoration
Danrom_'s Avatar

Danrom_

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.

Jellyfish haircut

Usage:

  • Lora will work even without triggers, but i strongly recommend you to use the trigger word "jfhaircut"

  • There is no exact phrase that you need to use, but for more control you should use something similar to this structure: "a woman with a long black hair with a jfhaircut" or "a woman with a long jfhaircut". Try to always define if you want long, short or mid-sized hair, also i recommend you to always specify the color that you want for the hair.

  • This lora can also produce a special type of jellyfish hair with braids, just use something like: "a woman with a short jfhaircut, her hair is [color] with braids on the sides".

Other:

  • This model was trained on synthetic data

  • This model was trained using flux 9b - Base and was tested with inference on Flux 9b - Distilled

  • The model was made for image generation, but it works in edit mode

  • Lora seems to preserve identity in edit, but i don't guarantee that

  • The model often fails to generate hair correctly in edit (about 1 time in 10-15 gens for me), but changing the seed is enough to solve it

My settings:

  • Flux Klein 9B

  • Sampler: Euler or Euler_a

  • Scheduler: Flux Scheduler

  • Cfg: 1.0

  • Steps: 4-12

  • Resolution range: 1024p - 1920p

  • Lora Strength: 1.0

Hope you like it. If you have any issue with this model, please let me know so i can improve it in further versions.