search

CMU-Perceptual-Computing-Lab / openpose

OpenPose: Real-time multi-person keypoint detection library for body, face, hands, and foot estimation
https://cmu-perceptual-computing-lab.github.io/openpose
Other
30.42k stars 7.8k forks source link

Smaller models #113

Closed mxmxlwlw closed 7 years ago

mxmxlwlw commented 7 years ago

Hi, I hope to get smaller models. Can you provide training codes? Or can you give me some new models? Especially those with small numbers of layers. I mean those basic conv layers which extract the foundational features. eg: conv1_1, conv1_2.... It seems they are greatly affect the speed of the network.

gineshidalgo99 commented 7 years ago

Training: https://github.com/ZheC/Realtime_Multi-Person_Pose_Estimation