Control-Net Open Pose Xinsir is not working in SD.Next Web UI
Hi , I encounter this in SD.Next Web UI where it is not working at all.
Here is my parameters:
Prompt: Male model posing in the street
Ready | Time 9.01s | Images 1 | Size 168x384
Prompt: Male model posing in the street
Parameters: Steps: 40| Seed: 1444209091| Sampler: Default| CFG scale: 6| Size: 168x384| Parser: Full parser| Model: sd_xl_base_1.0_0.9vae| Model hash: e6bb9ea85b| Backend: Diffusers| App: SD.Next| Version: 94f6f0d| Operations: control; txt2img| Pipeline: StableDiffusionXLImg2ImgPipeline| Control mode: ControlNet| Control conditioning: 1.0| Control start: 0.0| Control end: 1.0| Control model: Xinsir OpenPose XL| Size-1: 168| Size-2: 384
I tried with various combination of Pre-processor and ControlNet setting (including with DW Pose) , various Control conditions parameter but somehow it wont work.
May O know which WebUI have you tested this on ?
talk to vlad, the client needs a patch. refer to the comfy repo if you want to port his.
Sorry, I haven't do the Adaptation work to comfyui or ControlNet A1111 or diffusers, because my main energy lies training the model. The model is a new architecture and can't be directly used in Comfyui, when it comes to openpose, it should replace the draw_pose function as I have write in https://github.com/xinsir6/ControlNetPlus/tree/main?tab=readme-ov-file#inference-scripts
Sorry, I haven't do the Adaptation work to comfyui or ControlNet A1111 or diffusers, because my main energy lies training the model. The model is a new architecture and can't be directly used in Comfyui, when it comes to openpose, it should replace the draw_pose function as I have write in https://github.com/xinsir6/ControlNetPlus/tree/main?tab=readme-ov-file#inference-scripts
Thank you Xinsir for your hardwork. Thought that the models got some fault, but actually, the fault lies with the WebUI.