ValueError: Imaginary component 4.082076360939105e+125

[khansharkhamnida]

Hello, I've encountered the problem above

[vinesmsuic]

scipy has to be 1.11.1 instead of 1.11.2 . Just downgrade it with:

pip install scipy==1.11.1

[zc1023]

scipy has to be 1.11.1 instead of 1.11.2 . Just downgrade it with:

pip install scipy==1.11.1

Thank you！It works！

[1LOVESJohnny]

scipy has to be 1.11.1 instead of 1.11.2 . Just downgrade it with:

pip install scipy==1.11.1

It saves my day :)

[ElysionHuang]

scipy has to be 1.11.1 instead of 1.11.2 . Just downgrade it with:

pip install scipy==1.11.1

It does work lol !

[lucascolley]

It looks like the behaviour change comes from scipy/scipy#17918 in 1.11.2. That was supposed to be a bugfix - does it look like a regression here, or is this repo just relying on the buggy implementation?

[sh190128]

scipy has to be 1.11.1 instead of 1.11.2 . Just downgrade it with:

pip install scipy==1.11.1

niubi！

[wxc522]

scipy has to be 1.11.1 instead of 1.11.2 . Just downgrade it with:

pip install scipy==1.11.1

666

[lucascolley]

This should be fixed now in SciPy 1.13.0, which was just released. Please try with that SciPy version and let us know if the issue persists.

[davystrong]

@lucascolley just to confirm, this seems to be fixed now.

[lucascolley]

thanks @davystrong !

[lucascolley]

@mseitzer I recommend closing this issue now. You may want to consider:

• placing a lower bound of 1.13.0 on SciPy
• blocking scipy versions >=1.11.2, <1.13.0

Or just making the recommended fix of upgrading SciPy more visible.

[adil2414]

3 hr of looking for solution. it worked omagaaaa

[ShoufaChen]

scipy==1.13.0
numpy==1.26.2

works for me. numpy==1.23.5 does NOT work.

[ChongjianGE]

scipy==1.13.0
numpy==1.26.2

works for me. numpy==1.23.5 does NOT work.

@ShoufaChen Thanks! It saves my day!