search

NVIDIA / DIGITS

Deep Learning GPU Training System
https://developer.nvidia.com/digits
BSD 3-Clause "New" or "Revised" License
4.12k stars 1.38k forks source link

CAFFE_ROOT Problem ? #2032

Open earcz opened 6 years ago

earcz commented 6 years ago

Hi,

Note: (For a long time, I am looking for the solution on the internet and asked on devtalk.nvidia but found no solution)

First, I would like to say that in the beginning of installing DIGITS, I installed CUDA-9.1, cuDNN-7.1, Nvidia-Driver 390.48 and so on. I mean everything was the most updated and I followed the guide in NVIDIA/DIGITS. However, I thought that I should have installed older versions and did it in that way. And now, I installed DIGITS-6, Cuda-8.0 and cuDNN-6 regarding the installation guide of a NVIDIA's partner in Turkey. This is their installation guide website: (https://embedded.openzeka.com/ubuntu-16-04-cuda-cudnn-caffe-torch-jupyter-digits-kurulumu/)

My GNU nano 2.5.3 File: /home/deep/.bashrc export LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}} export PATH=/usr/local/cuda-8.0/bin${PATH:+:${PATH}} export CAFFE_ROOT=~/caffe export 'PYTHONPATH=/home/deep/caffe/python:${PYTHONPATH:+:${PYTHONPATH}} ' export DIGITS_ROOT=~/digits export TORCH_ROOT=~/torch . /home/deep/torch/install/bin/torch-activate

When I run ./digits-devserver

"/home/deep/caffe" from CAFFE_ROOT does not point to a valid installation of Caffe. Use the envvar CAFFE_ROOT to indicate a valid installation. Traceback (most recent call last): File "/usr/lib/python2.7/runpy.py", line 174, in _run_module_as_main "main", fname, loader, pkg_name) File "/usr/lib/python2.7/runpy.py", line 72, in _run_code exec code in run_globals File "/home/deep/digits/digits/main.py", line 70, in main() File "/home/deep/digits/digits/main.py", line 53, in main import digits.config File "digits/config/init.py", line 7, in from . import ( # noqa File "digits/config/caffe.py", line 226, in executable, version, flavor = load_from_envvar('CAFFE_ROOT') File "digits/config/caffe.py", line 37, in load_from_envvar import_pycaffe(python_dir) File "digits/config/caffe.py", line 126, in import_pycaffe import caffe File "/home/deep/caffe/python/caffe/init.py", line 1, in from .pycaffe import Net, SGDSolver, NesterovSolver, AdaGradSolver, RMSPropSolver, \ File "/home/deep/caffe/python/caffe/pycaffe.py", line 15, in import caffe.io File "/home/deep/caffe/python/caffe/io.py", line 8, in from caffe.proto import caffe_pb2 File "/home/deep/caffe/python/caffe/proto/caffe_pb2.py", line 1402, in options=_descriptor._ParseOptions(descriptor_pb2.FieldOptions(), _b('\020\001')), file=DESCRIPTOR), TypeError: new() got an unexpected keyword argument 'file'

I am aware of that it is a path problem, however, I could not fix this issue. I kindly request your interest on my problem. Thanks in advance.

Ender.

earcz commented 6 years ago

I solved the problem. I am not a professional but the problem was most probably the wrong version of installations. Here are what I installed: 1) Cuda 9.0 2) Cudnnv7.1 for cuda 9.0 3) Caffe 0.15.14

Note: I would not brave to install Tensorflow since it seemed a little tricky to me. Because I do not know which version of tensorflow should be installed although I checked for it on the official website of tensorflow. If anyone can help about this issue, I would be appreciate.