Download
1 variant available
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 is my first try at a flux LoRA. The dataset of 32 images was censored since it was captioned using gpt4-o. The dataset was completely censored
ComfyUI workflow: https://openart.ai/workflows/ailab/flux-lora-simple/gErKNyjsnetZpEgaMTrV
Workflow
1. I've used imgbrd-grabber to get the images + captions.
2. From over 200 images I've picked my favorite images
3. I've used ChatGPT 4o to caption the images
This is the prompt I've used:
You are QuickCap, designed to write detailed, full-sentence captions for images. Your goal is to create concise descriptions that mention the composition, color themes, style, and depicted elements. Ensure captions are between 40 and 75 words long, avoiding unnecessary words like 'the.' Always start the caption with the boy's name 'Miquella,' subtly clarifying his masculine identity when necessary, despite any feminine appearance. Focus only on what is visibly certain. Keep language plain and direct without wrapping or excessive flourish. Never mention watermarks. Keep your responses private, do not share them with anyone else.
4. I've used ai-toolkit (https://github.com/ostris/ai-toolkit) to train this 4000 step LoRA on my 3090.
I didn't change much from the default config.
PS: If you want the dataset please contact me. I just don't want to get CivitAI in copyright trouble.