diStyApps
commented
3 days ago Thank you very much, these are not nitty pick things it is very good feedback and it helps me a lot.
Some issues i am aware of, the "failed the load image" issue because it saved in a sub folder, this will be fixed soon, setting to be save in "output" folder should fix the issue.
For selecting Lora from drop-down you have to link manually there is no preset for this node, i will add that soon. like screenshot to see if that is what you mean and how it is been set.
Yes some custom nodes or some normal ComfyUI functionality will not work on API mode as normal mode, like some values gets hardcoded like time format, i need to add support from my end, so i will improve it as much as i can. Reporting on such issues will help me very much.
Tihf
Is there an existing issue for this?
Summary
Metadata with lora info for civitai
Description
So, i have this simple workflow, where it generates an image, refines it with face detailer from impact pack, and then send the image to Image Saver Node. https://github.com/alexopus/ComfyUI-Image-Saver
I use KJNodes Widget to String, that is able to get any value from any node by inputing it's ID and field name, usefull to send it to image saver https://github.com/kijai/ComfyUI-KJNodes
When i export the api e try running on flow i get this error: ComfyUI-KJNodes\nodes\nodes.py", line 833, in get_widget_value workflow = extra_pnginfo["workflow"] TypeError: 'NoneType' object is not subscriptable
which is normal, there's like a bunch of custom nodes out there, it's expected some of them would not work right outside the box
So, trying the workflow example of Image Saver, it was able to generate a image, but when the image is sent to face detailer,it keeps the blurred version of the image in the preview, when face detailer finishes, flow can not fetch the improved image and it turns to blank with the "failed the load image" error (but the image is saved anyhow in the output folder) the image loader component seems to not able to fetch the imagem from the image preview node aswell (but that's probably me not setting up correctly on flow)
and to able to get the lora info used for the image metadata, i need to use the specific lora node of the workflow, since the add lora button for flow will add lora in other ways that will not be connected with the image saver, but when i pick the lora input field from Lora Stacker Node (setting it up as a dropdown menu), it doesn't do anything, same with ImpactWildcardEncode lora field.
these are pretty much nitty pick things, so you don't need to bother about it too much, but this is nice addon for comfyui, and i thought if sharing it, would help improve the development somehow so, i will leave my workflow here if you want to test it
Eff Workflow with Face Detailer Metadata ImgSaver.json
Additional information
No response