It's not broken — the developers configured it this way on purpose. When you try to modify it, a message appears explaining that this setting is now in place to ensure the output images are correct and in optimal condition.
Rose Milkyway
•
Nov 1, 2025
I don’t think that benefits anyone. Seems like a very shifty decision Venice made against its users. Possibly for some legal, perhaps licensing, reason…
Jack
Team•
Nov 3, 2025
We made this change since the current Qwen-Image model is using a LORA. That LORA only works with 1 CFG and either 6 pr 7 on the steps value.
Rose Milkyway
•
Nov 30, 2025
Gotcha. Would be best explained in UI than in the comments of an issues board.
Log in to comment and vote
Comments4
Copper Honey
Sep 21, 2025
It's not broken — the developers configured it this way on purpose. When you try to modify it, a message appears explaining that this setting is now in place to ensure the output images are correct and in optimal condition.
Rose Milkyway
Nov 1, 2025
I don’t think that benefits anyone. Seems like a very shifty decision Venice made against its users. Possibly for some legal, perhaps licensing, reason…
Jack
Nov 3, 2025
We made this change since the current Qwen-Image model is using a LORA. That LORA only works with 1 CFG and either 6 pr 7 on the steps value.
Rose Milkyway
Nov 30, 2025
Gotcha. Would be best explained in UI than in the comments of an issues board.