Me, too. Most other sites have a download button, BUT . . . open Notepad, copy all the gobblety gook and paste it in Notepad. Save it as a .json file (but I think you'll have to delete the .txt afterwards, put it in your workflow folder and it should be good to go. But they did say it was already in the "Image" templates so you don't have to worry about that process I mentioned. But it's good to know in case you run into again. Merry Christmas!
I don't think it was that big. Didn't look like gigabytes. That "txt" file should be within the 27 kilobyte range. A model would be in the 10-20GB range. I just made the conversion and tested in Comfy. It's the workflow. Just don't forget to remove the ".txt" so it reads . . .json!
"convolution_overrideable not implemented. You are likely triggering this with tensor backend other than CPU/CUDA/MKLDNN, if this is intended, please use TORCH_LIBRARY_IMPL to override this function"
My question for the Comfy Pros here is what should/could I (we) do with all the 2509 models stuff? Does 2511 TAKE ITS PLACE? I guess good practice would be to keep 2509 for a while to make sure 2511 works well. I'd just like to hear the advice from the more knowledgeable before I delete stuff. I'm sure vaes and clip stuff are staying the same, right?
Wow! Great!
But the link which was supposed to be workflow (json file) looked to be way way larger than that.
That's the code behind the curtain of all computer programs. We just don't see it with the GUI!
Me, too. Most other sites have a download button, BUT . . . open Notepad, copy all the gobblety gook and paste it in Notepad. Save it as a .json file (but I think you'll have to delete the .txt afterwards, put it in your workflow folder and it should be good to go. But they did say it was already in the "Image" templates so you don't have to worry about that process I mentioned. But it's good to know in case you run into again. Merry Christmas!
It looked more like a model than a workflow...
I don't think it was that big. Didn't look like gigabytes. That "txt" file should be within the 27 kilobyte range. A model would be in the 10-20GB range. I just made the conversion and tested in Comfy. It's the workflow. Just don't forget to remove the ".txt" so it reads . . .json!
I am running Comfyui on my Mac Studio.
The Qwen Image Edit 2509 workflow ran fine.
However with this workflow I get:
"convolution_overrideable not implemented. You are likely triggering this with tensor backend other than CPU/CUDA/MKLDNN, if this is intended, please use TORCH_LIBRARY_IMPL to override this function"
Doesn't this one run on macOS?
Where can we find a list of the Chinese prompt words?
My question for the Comfy Pros here is what should/could I (we) do with all the 2509 models stuff? Does 2511 TAKE ITS PLACE? I guess good practice would be to keep 2509 for a while to make sure 2511 works well. I'd just like to hear the advice from the more knowledgeable before I delete stuff. I'm sure vaes and clip stuff are staying the same, right?
Yes, vaes and clips are the same.