Options to increase fps

[carstenschwede]

Are there any options to increase fps besides reducing resolution or adding GPUs? Is it possible to restrict detection to certain joints (e.g. Heads) in order to speed up processing?

[ZheC]

(1) Using MPI model instead of COCO model (2) Using one scale for testing can speed up the processing time. Restricting the detections does not help because the CNN still need to use the same trained model and thus the CNN forwarding processing time is the same.

[ZheC]

Another option is to modify the text.prototxt and reduce the stage number from 6 to 3.

[carstenschwede]

Thanks, I will try both. Any idea of what kind of speedup I could expect?

[Warden7]

Hi I have a question about the fps is that: I run the rtpose demo on the AWS p2.large instance(with one K80 GPU 24G), however it takes 1.1s to deal a frame.
I don't know whether it is because that the k80 gpu has a compute capability of 3.7 lower than that of 6.1 of GTX1080?

[gineshidalgo99]

These is a preliminary benchmark we have made with the new version we are working on (it will be released in around 1 month). The current version you are using should be around 25-30% slower. Let me know if you are using the same flags. If so, are you using cuDNN 5.1? Older versions of cuDNN might also slow down the program. Thanks!

Current benchmark:
https://docs.google.com/spreadsheets/d/1-DynFGvoScvfWDA1P4jDInCkbD4lg0IKOYbXgEq0sK0/edit#gid=0

[wangzhangup]

@Warden7
their compute capabilities
K80: 8.73TFLOPS
1080: 9TFLOPS

low fps maybe other reasons

[Warden7]

Thanks for your warmly analysis. The version of cuDNN is 5.0 and Cuda is 7.5. The key word of GPU information "volatile gpu util" always shows 99%, even though nothing is done on the GPU.Maybe something debug need to be done further.

[wangzhangup]

@Warden7
kill the processes on the GPU

[gineshidalgo99]

Another way to speed it up is by using the new version (~25% faster):
https://github.com/CMU-Perceptual-Computing-Lab/openpose

[wangzhangup]

Reduce the number of feature maps.
I modify the stage 3-6 conv layer's output number from 128 to 64. And the result is as good as original version, speed up 25%!

[carstenschwede]

@wangzhangup Thanks, can you try your modification also on the newer version at https://github.com/CMU-Perceptual-Computing-Lab/openpose? Would be interesting to see what overall speedup you are able to get.

[carstenschwede]

@gineshidalgo99 Thanks for the update!

[gineshidalgo99]

@wangzhangup Thank you so much for your idea! Please, could you email me: gines@cmu.edu to discuss how you did it in more details? We are interested in adding it to our system if that is OK for you!