I am working on caffe2. I am running Densepose project from FaceBook research. By running the inference code, I get these warning
WARNING cnn.py: 25: [====DEPRECATE WARNING====]: you are creating an object from CNNModelHelper class which will be deprecated soon. Please use ModelHelper object with brew module. For more information, please refer to caffe2.ai and python/brew.py, python/brew_test.py for more information.
INFO net.py: 51: Loading weights from: /tmp/detectron-download-cache/DensePose_ResNet101_FPN_s1x-e2e.pkl
I1128 12:13:25.745167 22255 operator.cc:195] Engine CUDNN is not available for operator MaxPool.
I1128 12:13:25.750291 22255 operator.cc:195] Engine CUDNN is not available for operator MaxPool.
I1128 12:13:25.751339 22255 net_dag_utils.cc:102] Operator graph pruning prior to chain compute took: 6.9125e-05 secs
I1128 12:13:25.754218 22255 operator.cc:195] Engine CUDNN is not available for operator MaxPool.
I1128 12:13:25.759892 22255 operator.cc:195] Engine CUDNN is not available for operator MaxPool.
I1128 12:13:25.760098 22255 net_dag_utils.cc:102] Operator graph pruning prior to chain compute took: 5.4858e-05 secs
I1128 12:13:25.761886 22255 net_dag_utils.cc:102] Operator graph pruning prior to chain compute took: 1.3314e-05 secs
I1128 12:13:27.910995 22255 net_async_base.h:201] Using specified CPU pool size: 4; device id: -1
I1128 12:13:27.911011 22255 net_async_base.h:206] Created new CPU pool, size: 4; device id: -1
I1128 12:13:28.632196 22255 net_async_base.h:201] Using specified CPU pool size: 4; device id: -1
I1128 12:13:28.632215 22255 net_async_base.h:206] Created new CPU pool, size: 4; device id: -1
My problem is I had a very low frame rate (5-7 fps) compared to the original paper (20-26). I am wondering if the frame rate problem is coming from these MaxPool which probably done on CPU. Any help?