Open franzkruhm opened 6 months ago
Is it a standard or custom setup for the prompts?
I experienced the same wrong behaviour. A workaround is to create an extra primitive string nodes for the positive/negative prompts and connect them to the proper Clip Text encode positive/negative and also to the Save image extended positive_text_opt/negative_text_opt. Then it works, but it's quite additional work.
Is it a standard or custom setup for the prompts?
I just created a new workflow by clicking to the Load Default. So imho it's standart setup for the prompt.
Hi, this is a bug that I am also experiencing. The JSON file includes a field for the "positive prompt" but it actually includes the negative prompt. I need both of them to be in the JSON file, and correctly labeled.
I was able to work around it by adding extra text nodes to drive the conditioning and the inputs on Save Image Extended. But this is almost certainly a simple fix in the python code. It would be great if this could be dealt with. Thanks!!
Hello Everyone,
I had the same issue and fixed the code. I found a typo bug in Line 239 in save_image_extended.py. Basically the positive prompt was overwritten by the negative one. You can fix it yourself by correcting the line until it's fixed :)
Just replace the second "positive_prompt" with "negative_prompt".
@thedyze @AlUlkesh
I had the same issue and fixed the code.
Awesome, thanks. I knew it would be an easy fix.
@addiGeometry please submit your PR here, thanks
@audioscavenger Hey Eric. This is my first time contributing on github. I tried creating the pull request. let me know if it worked. Weirdly enough git seems to see way more changes then there actually were...
So I have![image](https://github.com/thedyze/save-image-extended-comfyui/assets/34811816/101b02d3-e0cb-4df5-85ac-c8eab7323f9f)
but I get just one "positive prompt" field in the json for each pic but it contains... the negative prompt