Re:PP-YOLO Surpasses YOLOv4 – State-of-the-art object detect...
by Imnimo on Monday August 03, 2020 @09:10PM
(#24045455)
I think at a certain point, FLOP count will be more important than FPS. Like once you're running at real time, there aren't a lot of applications that care about 120 FPS vs 110 FPS. But there are a lot of situations where you care about the total number of operations (regardless of GPU parallelism) because you want to run on an edge device or have power constraints.
[ Parent
| on HN ]
Re:PP-YOLO Surpasses YOLOv4 – State-of-the-art object detect...
by CompleteSkeptic on Monday August 03, 2020 @09:29PM
(#24045563)
There actually is some work (https://arxiv.org/abs/2003.13630) claiming that FLOPS are a poor measure of real-world performance - with some of the more recent FLOP-efficient models actually running slower than older models.
Re:PP-YOLO Surpasses YOLOv4 – State-of-the-art object detect...
by flafla2 on Tuesday August 04, 2020 @07:49AM
(#24048986)
Forgive me if I’m being dense, but shouldn’t we expect performance to degrade if flop count per unit time is decreased, assuming performance is defined as overall runtime (FPS in this case)? It’s a trade off scenario where runtime performance is being balanced by other concerns such as power consumption.
Re:PP-YOLO Surpasses YOLOv4 – State-of-the-art object detect...
by jimmySixDOF on Tuesday August 04, 2020 @01:13AM
(#24046865)
>because you want to run on an edge device or have power constraints.
Plug for TinyML
Re:PP-YOLO Surpasses YOLOv4 – State-of-the-art object detect...
by ganstyles on Monday August 03, 2020 @09:33PM
(#24045587)
We have started to hit this in the last two years. Models get deeper, more FLOPS but not practically better metrics except FPS. But for our use cases and on our hardware constraints, 30fps on a model that's a couple years old outperforms even the Yolov99 or whatever is SOTA.