Verified: 10 months 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.
Please Donate Buzz for FLUX Lora Training !
FLUX v1.0
Lora trained on 80 images on the concept "tongue out" and "long tongue out". the captioning was done with GPT-4 Vision on FLUX 1.D. Tested on FLUX 1.D fp8, NF4 and FLUX 1.D [full]. The strength should be 0.8-1.2. Depending on the type of image you can specify keywords as follows. Main keyword is "{someone} sticking tongue out". Some additional ones are:
View keywords : "Frontview", "Sideview"
Tongue length : "Long tongue out", "tongue out"
Mouth openness : "Mouth wide open, uvula"
So a positive prompt could be something like this:
Positive : Frontview, mouth wide open, {style and scene description}, {character description} with {her/his/their} long tongue sticking out,<lora:concept_long_tongue_out_flux_1_standard-000017:1>
Keep the Lora strength at around 1.0. It can be used with other Loras!
SDXL v1.0
Lora trained on 83 images on the concept "tongue out" and "long tongue out". the captioning was done with GPT-4 Vision. The strength should be 0.4-0.9. Depending on the type of image you can specify keywords as follows. Main keyword is "{someone} sticking tongue out". Some additional ones are:
View keywords : "Frontview", "Sideview"
Tongue length : "Long tongue out", "tongue out"
Mouth openness : "Mouth wide open, uvula"
If you have a specific non-realistic style you want to adhere to then keep strength around 0.4-0.8 for better prompt adherence.