np.float errors in the trackers

[thecetin]

Hello, im using trackers algorithm in "colab". So i always need to install in to "colab" when i work on some projects. So every time i need to edit tracker code and replace all "np.float" to "float". How can this problem solve?

->ByteTrack -> yolox -> tracker -> byte_tracker.py and matching.py

In these two .py files you should change "np.float" to "float"

Thank you for help.

[braiansmarzaro]

I was about to open this issue. I'm having the same problem.

[LemonJam84]

I am also having issues here, have you @braiansmarzaro and @thecetin any luck solving the problem?

[braiansmarzaro]

Hi @LemonJam84. As far as i can remember, i cloned the repo and changed every np.float to just float and it worked. Good luck with this problem 🫡

[LemonJam84]

sounds good, thanks @braiansmarzaro

[iurisilvio]

The quick fix is limiting numpy version <1.24, but it is a numpy version released in 2022.

There are a few open PRs ~possibly~ fixing this error:

• #399
• #360
• #389
• #396

And related issues:

• #311
• #330

[NicDionne]

@ifzhang This would be very helpful for keeping the library up to date.