Closed 2p990i9hpral closed 5 years ago
Maybe you can try the smaller model like SqueezeNet and read frames in batch, the current implementation is 1 batch size, which will cause the low gpu usage.
Maybe you can try the smaller model like SqueezeNet and read frames in batch, the current implementation is 1 batch size, which will cause the low gpu usage.
also i want skip uncertain pose like "https://ibb.co/C1VCX0r" can I skip uncertain pose on code?
you can extract the pose score from heatmap and then delete the low-score keypoint when visualization.
you can extract the pose score from heatmap and then delete the low-score keypoint when visualization.
thank you for your answer this project is very useful
28
After completing the steps above, gpu usage is very low it this a cpu bottleneck? i want a extreme speed Beyond real-time
on my gpu(1080ti) it's only 45 fps. gpu usage is 4~5%, vram usage is 500mb