Black Forest Labs’ (BFL) FLUX.2 is now available on Microsoft Foundry. Building on FLUX1.1 [pro] and FLUX.1 Kontext [pro], we’re excited to introduce FLUX.2 [pro] which continues to push the fronti...
As someone who is just getting started, I would like to sincerely thank you for the thorough overview of Flux2‑pro’s strengths, which has been very helpful in deepening my understanding of the FLUX2 model.
However, when testing Flux2-pro in Azure AI Foundry, I noticed that neither the example code nor the Playground seems to provide guidance or samples on how to perform image‑based(image to image) editing. Additionally, the API invocation pattern for Flux2 has changed significantly, which has introduced some challenges for my current debugging efforts.
FLUX1-Kontext:
FLUX2-Pro:
My longer‑term objective is to apply this capability to a virtual try‑on use case, which involves multiple image inputs and aspires to achieve improved results compared to FLUX.1‑Kontext‑pro. With that in mind, I was wondering whether Azure AI Foundry might consider offering more complete, step‑by‑step guidance, similar to what exists for FLUX.1‑Kontext‑pro, especially for image editing scenarios.
Thank you very much for your time, and I look forward to hearing your thoughts