Skip to content
This repository has been archived by the owner on Jun 15, 2022. It is now read-only.

Improvement speed #31

Open
arndey opened this issue May 6, 2018 · 4 comments
Open

Improvement speed #31

arndey opened this issue May 6, 2018 · 4 comments

Comments

@arndey
Copy link

arndey commented May 6, 2018

Hi! On converted caffe model estimation accurate is so good, but speed is slow for me. I wrote script for runing estimation on video like camera_pose_demo.py and with gtx1050 it shows 2fps. Can I somehow improve speed without accuracy worsing? Maybe somebody can give his owntrained model, or give me some recomendations? Thanks!

@arndey arndey changed the title Improving speed Improvement speed May 6, 2018
@arndey
Copy link
Author

arndey commented May 6, 2018

And I need only single person estimation.

@GhadeerMohamad
Copy link

Hello, i trained the model for keypoints detection only(without pafs), i can pass the pretrained model to you if that would help.

@arndey
Copy link
Author

arndey commented Jun 5, 2018

@GhadeerMohamad yes, it would be great. You can get my email in bio

@GhadeerMohamad
Copy link

Good luck, and don't forget that the file doesn't use pafs so it wouldn't work if you don't remove the pafs branch first.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants