daniandtheweb/sd.cpp-webui

AI Summary: The user is requesting the addition of support for 'Wan 2.1 and 2.2' text-to-video models within the existing web UI. This enhancement aims to integrate these newer model versions, making them available for users.

AI Summary: The UNET generation page in the web UI incorrectly passes an empty `-n` (negative prompt) argument to the `sd.cpp` command, which causes a parsing error when using `--diffusion-model` paths. This issue does not occur when using `--model` (checkpoint) paths with the same empty `-n`. The user requests adding a negative prompt field to the UNET page and ensuring `""` is always passed for `-n` if no negative prompt is specified.

AI Summary: Debug a Gradio-based Stable Diffusion web UI. The issue is an 'invalid mode txt2img' error when using the latest version of stable-diffusion.cpp. The task involves investigating why the web UI is not correctly passing the 'txt2img' mode to the underlying stable-diffusion.cpp executable, potentially requiring examination of command-line argument handling within the Python script and/or the stable-diffusion.cpp API.