Query regarding gpu and cpu usage

[Hemanth-TS]

when i ran the command with cuda , around 20% of gpu was used and like 200% cpu but it was still very slow, i was trying to infer a video.
python yolo/lazy.py task=inference \ # default is inference
name=AnyNameYouWant \ # AnyNameYouWant
device=cuda\ # hardware cuda, cpu, mps
model=v9-s \ # model version: v9-c, m, s
task.nms.min_confidence=0.1 \ # nms config
task.fast_inference=onnx \ # onnx, trt, deploy
task.data.source=data/toy/images/train \ # file, dir, webcam
+quite=True \ # Quite Output

Can someone guide me on how to infer videos properly with higher fps while using gpu efficiently.

[henrytsui000]

Hi,

Could you please provide your system details, command line, and Git commit version? For example, my setup is:

• Git Tag/Branch: main (commit add9e2f)
• OS: Ubuntu 22.04
• GPU: NVIDIA RTX 3090
• CPU: AMD Ryzen 9 3900X
  Command:

python yolo/lazy.py task=inference name=AnyNameYouWant device=cuda model=v9-c task.data.source=MOT20-05-raw.mp4 task.fast_inference=deploy

This configuration achieves 20 FPS without using ONNX or TensorRT.

best regards,
Henry Tsui