Is the code runnable on Raspberry Pi?

gaomingqi / Track-Anything

Track-Anything is a flexible and interactive tool for video object tracking and segmentation, based on Segment Anything, XMem, and E2FGVI.

MIT License

6.45k stars 480 forks source link

Is the code runnable on Raspberry Pi? #40

Open franva opened 1 year ago

franva commented 1 year ago

First of all, you guys did an amazing job 👏💯💯💯👏!!

I've always wanted to track objects in videos found other projects which can do pretty good job for tracking, but not able to run on Arm based edge devices e.g. RPi.

So just wondering is TAM for tracking purpose able to be run on Raspberry Pi 4(Mine has 4GB memory)?

Looking forward to your reply.

Cheers, Winston

gaomingqi commented 1 year ago

Hello, thanks for your interest! :raised_hands: We have not tested TAM on Raspberry Pi. I think memory seems ok since TAM can be configured more lightweight. But I am not sure RPi supports the required environments. Thanks.

franva commented 1 year ago

@gaomingqi Thanks for your reply. Do you have any plan to test it on RPi 4? Being able to run on the edge devices like RPi will be a big benefit for TAM. I have RPi 4 with 4GB RAM, I'm happy to test it, but it seems TAM requires GPU and cuda libraries? How could we make TAM more self-contained?

Cheers,

Winston