SlipperySnekySneks
commented
1 week ago To add to this, minimizing the browser, clicking into another window (not the desktop, it has to be another window), then going back to WebUI will fix the problem, albeit temporarily. When the issue occurs, clicking on the image causes it to switch to the next image in the set, rather than loading the full size image.
As stated in the title, in Forge UI, upon start-up, primarily running flux models, the gallery will not fullscreen the image until a few generations have completed. The gallery isn't unresponsive; I can close out the image to the thumbnail view, send the gen to img2img, and even save it. I have confirmed this behavior across multiple devices (Android, iPad, Windows) and through various browsers: Chrome and Firefox. I see no errors or issues in the console.
I've tried disabling arguments like: cuda malloc, cuda stream. To see if maybe those were some of the issues.
Honestly, its not a huge bug, but its a minor annoyance having to gen 4-5 images before I can use the UI as it always has before.
Specs: AMD 5950X 64G DDR4 RTX4090 Windows 11 Firefox: 130.0 (64bit)
Startup:
Python 3.10.11 (tags/v3.10.11:7d4cc5a, Apr 5 2023, 00:38:17) [MSC v.1929 64 bit (AMD64)] Version: f2.0.1v1.10.1-previous-459-gd3396001 Commit hash: d339600181252aafbe12dc06f1b4fa7f953fc6f3 Launching Web UI with arguments: --listen --enable-insecure-extension-access --cuda-malloc --cuda-stream Using cudaMallocAsync backend. Total VRAM 24564 MB, total RAM 65459 MB pytorch version: 2.3.1+cu121 Set vram state to: NORMAL_VRAM Device: cuda:0 NVIDIA GeForce RTX 4090 : cudaMallocAsync VAE dtype preferences: [torch.bfloat16, torch.float32] -> torch.bfloat16 CUDA Using Stream: True Using pytorch cross attention Using pytorch attention for VAE Model selected: {'checkpoint_info': Using online LoRAs in FP16: True Running on local URL: http://0.0.0.0:7860