search

trrahul / densepose-video

Code to run densepose on video with detectron. https://github.com/facebookresearch/Detectron
GNU General Public License v3.0
62 stars 18 forks source link

Script hangs #13

Open pavithraezhil opened 5 years ago

pavithraezhil commented 5 years ago

Hello, I am unable to run the model even after using ResNet50 weights.

python2 tools/infer_vid.py --cfg configs/DensePose_ResNet50_FPN_s1x-e2e.yaml --output-dir DensePoseData/infer_out/ --wts DensePose_ResNet50_FPN_s1x-e2e.pkl --input-file test_video_hallway.mp4 Found Detectron ops lib: /usr/local/lib/libcaffe2_detectron_ops_gpu.so 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: DensePose_ResNet50_FPN_s1x-e2e.pkl [I net_dag_utils.cc:102] Operator graph pruning prior to chain compute took: 3.0053e-05 secs [I net_dag_utils.cc:102] Operator graph pruning prior to chain compute took: 2.2634e-05 secs [I net_dag_utils.cc:102] Operator graph pruning prior to chain compute took: 4.814e-06 secs |Processing Frame 1/276 -Frame read in 0.150s [I net_async_base.h:212] Using specified CPU pool size: 4; device id: -1 [I net_async_base.h:217] Created new CPU pool, size: 4; device id: -1 Killed

System Config: Ubuntu 16.04, CUDA 10 OpenCV 3.4.4, Caffe2 from source with GPU support GPU: 2080Ti, 11GB memory I am using python2.7 via virtualenv

Please help. Thank you.

trrahul commented 5 years ago

Are you sure your caffe installation is with GPU support?

pavithraezhil commented 5 years ago

Hi, Yes, I'm sure that caffe2 is compiled with GPU support. I have also executed the command on caffe2 to find the number of CUDA devices for GPU computation and it returns the right value. Also, while running the densepose code, when I monitor system memory, all of the GPU memory (11GB) and swap memory gets occupied. I am not sure how to go about solving this.