Invoke is a leading creative engine for Stable Diffusion models, empowering professionals, artists, and enthusiasts to generate and create visual media using the latest AI-driven technologies. The solution offers an industry leading WebUI, and serves as the foundation for multiple commercial products.
Summary
This PR adds support for the XLabs FLUX IP-Adapter v2 model (https://huggingface.co/XLabs-AI/flux-ip-adapter-v2).
Known v1 regression
The recommended inference configuration for XLabs IP v2 is different than that for v1:
This PR updates the Linear UI behavior to optimize for v2. As a result, the performance of v1 models in the Linear UI will be significantly degraded.
It is still possible to use the v1 models by manually constructing a workflow like the one shown on the original IP-Adapter v1 PR: https://github.com/invoke-ai/InvokeAI/pull/7157
Example 1:
a statue wearing sunglasses
In order (left to right, top to bottom):
Example 2:
A hooded warrior shown from behind overlooking a vibrant underwater sea world. There are sea turtles and colorful fish swimming around.
In order (left to right, top to bottom):
QA Instructions
Checklist
What's New
copy (if doing a release after this PR)