Is it not possible to use IP Adapter with DirectML?

[kofteburger]

I've tried to use the IP Adapter Controlnet Model with this port of the WebUI but it failed.
As I understand it, IP adapter uses Clipvisiondetector which only supports CUDA or CPU.
Would it be possible to force CPU only just for IP Adapter model?

RuntimeError: Attempting to deserialize object on a CUDA device but torch.cuda.is_available() is False. If you are running on a CPU-only machine, please use torch.load with map_location=torch.device('cpu') to map your storages to the CPU.

[kofteburger]

For those searching for this. I didn't find a way to get IP Adapter working with Automatic1111 DirectML it works on ComfyUI DirectML using this custom node. It runs without controlnet and the clip vision part is handled by the cpu.

[DrMacabre]

do you have a workflow to share? i tried with the one available on comfyUi ipadapter plus page but it's throwing errors and get stuck on the ksampler node.

[kofteburger]

I had since caved in and installed Linux but as I recall it worked with the default workflow available on the repo page.