caffe2 no gpu support error: libprotobuf.so.14 [In Progress]

Question

Currently running AWS Deep Learning AMI (Ubuntu) Version 17.0 and have run into a bump with Caffe2. Checking to see if caffe2 was correctly installed by using

python -c 'from caffe2.python import workspace; print(workspace.NumCudaDevices())'

I receive the following error.

Current Error:

ubuntu@ip:~$ python -c 'from caffe2.python import workspace; print(workspace.NumCudaDevices())'
WARNING:root:This caffe2 python run does not have GPU support. Will run in CPU only mode.
WARNING:root:Debug message: libprotobuf.so.14: cannot open shared object file: No such file or directory
CRITICAL:root:Cannot load caffe2.python. Error: libprotobuf.so.14: cannot open shared object file: No such file or directory

What am i missing? Not sure if this is helpful but it seems to exist.

File Found?

    ubuntu@ip:~$ find . -name "libprotobuf.so.14*"
./anaconda3/pkgs/protobuf-3.4.0-py27_1/lib/libprotobuf.so.14.0.0
./anaconda3/pkgs/protobuf-3.4.0-py27_1/lib/libprotobuf.so.14
./anaconda3/pkgs/protobuf-3.4.0-py35_1/lib/libprotobuf.so.14.0.0
./anaconda3/pkgs/protobuf-3.4.0-py35_1/lib/libprotobuf.so.14
./anaconda3/envs/caffe_p35/lib/libprotobuf.so.14.0.0
./anaconda3/envs/caffe_p35/lib/libprotobuf.so.14
./anaconda3/envs/caffe2_p27/lib/libprotobuf.so.14.0.0
./anaconda3/envs/caffe2_p27/lib/libprotobuf.so.14
./anaconda3/envs/caffe_p27/lib/libprotobuf.so.14.0.0
./anaconda3/envs/caffe_p27/lib/libprotobuf.so.14

AWS Instance: With latest deep learning frameworks pre-installed: MXNet, TensorFlow, PyTorch, Keras, Chainer, Caffe/2, Theano & CNTK, configured with NVIDIA CUDA, cuDNN, NCCL & Intel MKL-DNN.

Answer 1

Based on your search

It is found that "libprotobuf.so.14"

is available at ./anaconda3/envs/caffe_p35/lib/libprotobuf.so.14 [one of the path]

Try adding this lib path to LD_LIBRARY_PATH environment variable and Run the code again.

Below is the command to set the path from where you have executed the 'find' command

---

export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:./anaconda3/envs/caffe_p35/lib"

---

Hope this will solve your issue.