Download
1 variant available
144
117
1.2K
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.
Kontext has become my tool of choice for creating character LoRAs, and one of the biggest weaknesses I have noticed is profile views. The base Kontext models struggle to consistently generate full side/profile images of characters. This is not something I want to deal with every time I make a character, so this LoRA is a step towards a consistent solution!
V1 has a small dataset, so it does best with images of humanoid characters on simple backgrounds that focus on the head with as little shoulder/upper body as possible. You may also need to tell it to "preserve" important aspects of the image, like the image style (the dataset is entirely realistic) or the character's expression.
The trigger phrase for this LoRA is: Generate a profile headshot of the character.
Suggested strength: 0.75-1.25
Known Issues:
No way to choose direction
Complex backgrounds reduce the LoRA's effectiveness
Limited nose shapes in dataset
No non-humans in dataset
No cartoon/anime in dataset (minor issue, can be worked around with prompting)
Please leave questions/comments/suggestions/advice/etc. in the comments! I am working on V2 but wanted input from others and figured this would be the best way to do it!