Running on colab ..Unable to extract feature using c3d

[suyash8991]

extract feature

%cd /content/C3D/C3D-v1.0/examples/c3d_feature_extraction !sudo bash c3d_sport1m_feature_extraction_video.sh

This is the error I am getting while executing the above code. I0117 06:43:11.085855 4186 net.cpp:183] Collecting Learning Rate and Weight Decay. I0117 06:43:11.085873 4186 net.cpp:176] Network initialization done. I0117 06:43:11.085881 4186 net.cpp:177] Memory required for Data 5653289408 E0117 06:43:11.816212 4186 extract_image_features.cpp:81] Extracting features for 1 batches F0117 06:43:13.534600 4186 vol2col.cu:75] Check failed: error == cudaSuccess (48 vs. 0) no kernel image is available for execution on the device *** Check failure stack trace: *** c3d_sport1m_feature_extraction_video.sh: line 3: 4186 Aborted (core dumped) GLOG_logtosterr=1 ../../build/tools/extract_image_features.bin prototxt/c3d_sport1m_feature_extractor_video.prototxt conv3d_deepnetA_sport1m_iter_1900000 0 50 1 prototxt/output_list_video_prefix.txt fc7-1 fc6-1 prob

[aliakber366]

I ran into the same issue, @suyash8991 did you find any solution for this error?

[dolongbien]

Hi, make sure you were running under the GPU mode of Colab. Also, double-check the CUDA version and follow the right instruction on C3D notebook of this repo.

[mh-y]

Also, feature extractions for examples fail when already completed. Please make sure that the feature has already been extracted.

[harshith7823]

I too got something similar -> F0228 17:00:14.112013 6347 syncedmem.cpp:47] Check failed: error == cudaSuccess (46 vs. 0) all CUDA-capable devices are busy or unavailable If anyone knows how to resolve this, pls let me know

[ericbyeric]

Try the following commands in the beginning of the code

install cuda 10.2

!wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin !sudo mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600 !sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub !sudo add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/ /" !apt-get update !apt-get -y install cuda-10.2

I resolved the same issue in this way