Instructions to use XLabs-AI/flux-ip-adapter with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use XLabs-AI/flux-ip-adapter with Diffusers:
pip install -U diffusers transformers accelerate
from diffusers import ControlNetModel, StableDiffusionControlNetPipeline controlnet = ControlNetModel.from_pretrained("XLabs-AI/flux-ip-adapter") pipe = StableDiffusionControlNetPipeline.from_pretrained( "fill-in-base-model", controlnet=controlnet ) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
No Selection ?
#29
by DerD4nnY - opened
Hey there, the old IP Adapter always gave u the option to select if you want composition or style ... or both ... why is this missing for Flux ?
well, old sdxl ipadapter style or composition is achieve by different unet layer and multiple ways to do decouple cross attention.
flux is dit structure, the different layer(dit block) which affect style or composition is still unexamined.
in general, this ipadapter is working not as good as SDXL ones by far.
community needs more time to discover ways of adapters to affect the dit structure.