Do we support inference on video stream?

[Hui_Liu]

Hi,

I only see examples at Home | PyTorch on static images.

Do we have plan to support inference on video stream ( and automatically drop frame before inference finishes etc)?

I do see this: [Feat] Camera Stream API proposal · Issue #2920 · pytorch/vision · GitHub but just to reconfirm

Thanks!

[marksaroufim]

Sure ios-demo-app/TorchVideo at master · pytorch/ios-demo-app · GitHub

[Hui_Liu]

Thanks, Mark! Will see if I could accomodate that to do object detection directly on camera stream