search

CMU-Perceptual-Computing-Lab / caffe_rtpose

Realtime C++ code for multi-person pose estimation
Other
357 stars 207 forks source link

Real time C++ code for multi-person pose estimation,when make all,it appears some problems #2

Closed tanggt closed 7 years ago

tanggt commented 7 years ago

[platform] Ubuntu 14.04 +cuda8.0+cudnn5.1+opencv2.4.13 when I make all,it appears some problems: 1.std::runtime_error is not member of std,which could be solved by insert the header of stdexcept,then it will be make successfully! 2.After make all,when I run the command of "./build/examples/rtpose/rtpose.bin",it appear a error about threads, which is: E1230 09:52:39.072165 3477 rtpose.cpp:1507] Finish spawning 1 threads. now waiting. init done opengl support available However,then it all always waiting,no results,could anyone help me solver the problem.Thank you!

tsimk commented 7 years ago
  1. Should be fixed by 707e1aed9f817fd4dd824afc66c35a4566c203f8
  2. This is not an error, just an info message. I updated this in 53962cd48c0dd6239abe616555d136cf25731d09

As for not showing any results, I'm not sure what is happening. You may try running on an mp4 video instead, and adding the "--logtostderr" option to get more verbose logging of what is going on.

tanggt commented 7 years ago

I may find the key problem,which is the opencv version may be not compatible. The opencv version of mine is opencv-2.4.13.I compared it between laptop computer(opencv-2.4.9) and desktop computer(opencv-2.4.13).In laptop,the window could appear,however,the GPU memory is out.The desktop could not appear window,but using the command you suggest that could have results,which also to be add --no_display

./build/examples/rtpose/rtpose.bin --video test.mp4 --write_frames output/ --logtostderr --no_display I1230 19:28:11.457729 14254 rtpose.cpp:1711] Display resolution: 1280x720 I1230 19:28:11.457806 14254 rtpose.cpp:1717] Net resolution: 656x368 I1230 19:28:11.457844 14254 rtpose.cpp:1484] Finish spawning 1 threads. I1230 19:28:11.457875 14257 rtpose.cpp:175] Setting GPU 0 I1230 19:28:11.700845 14257 rtpose.cpp:180] GPU 0: copying to person net I1230 19:28:15.637430 14257 rtpose.cpp:202] start_scale = 1 I1230 19:28:15.637516 14257 rtpose.cpp:231] Dry running... I1230 19:28:15.907531 14257 rtpose.cpp:233] Success. I1230 19:28:15.907768 14257 rtpose.cpp:1081] GPU 0 is ready I1230 19:28:21.852883 14266 rtpose.cpp:1451] # 142, NP 6, Latency 0.274, Preprocess 0.000, QueueA 0.052, GPU 0.201, QueueB 0.003, Postproc 0.008, QueueC 0.000, Buffered 0.000, QueueD 0.010, FPS = 4.1 I1230 19:28:27.976358 14266 rtpose.cpp:1451] # 263, NP 4, Latency 0.289, Preprocess 0.000, QueueA 0.068, GPU 0.201, QueueB 0.003, Postproc 0.008, QueueC 0.000, Buffered 0.000, QueueD 0.009, FPS = 4.9 I1230 19:28:34.141719 14266 rtpose.cpp:1451] # 385, NP 4, Latency 0.312, Preprocess 0.000, QueueA 0.077, GPU 0.201, QueueB 0.005, Postproc 0.009, QueueC 0.010, Buffered 0.000, QueueD 0.010, FPS = 4.9 I1230 19:28:40.354110 14266 rtpose.cpp:1451] # 509, NP 3, Latency 0.289, Preprocess 0.000, QueueA 0.067, GPU 0.200, QueueB 0.003, Postproc 0.008, QueueC 0.000, Buffered 0.000, QueueD 0.010, FPS = 4.8 I1230 19:28:45.436137 14263 rtpose.cpp:526] Done, exiting. # frames: 615

tsimk commented 7 years ago
tanggt commented 7 years ago
  1. Desktop computer: gtx1060 6G.
  2. I Check the jpg files in output/ and the results look reasonable. I think it has something with opencv
gineshidalgo99 commented 7 years ago
  1. If you are running out of memory with the GTX 1060, the issue might be similar to #3, and you might need to enable cuDNN on the Makefile.config. Caffe CNN uses a higher GPU memory than the cuDNN implementation.
  2. Please, let us know if you were able to solve this OpenCV issue.