Dockerfile for pytorch in python2 won't build

[dbdean]

Hi, thanks for the help in getting pytorch running in python 2.7 using your Dockerfile at https://github.com/researchapps/sherlock/blob/master/pytorch-dev/Dockerfile.py2. I did have to make some changes to get it to work however:

+++ Dockerfile.new      2018-08-14 09:44:08.542990180 +1000
@@ -20,17 +20,19 @@
      /opt/conda/bin/conda config --add channels conda-forge && \
      /opt/conda/bin/conda install -y numpy pyyaml scipy ipython mkl && \
      /opt/conda/bin/conda install -y matplotlib scikit-learn scikit-image && \
+     /opt/conda/bin/conda install -y typing && \
      /opt/conda/bin/conda install -c intel mkl-include && \
      /opt/conda/bin/conda install -c pytorch magma-cuda90 && \
      /opt/conda/bin/conda clean -ya
-RUN /opt/conda/bin/conda update -y base conda
+RUN /opt/conda/bin/conda update -y conda
 ENV PATH /opt/conda/bin:$PATH
 # This must be done before pip so that requirements.txt is available
 RUN git clone https://github.com/pytorch/pytorch /opt/pytorch && \   
     cd /opt/pytorch && \
     git submodule update --init
 RUN pip install --upgrade pip
-RUN TORCH_CUDA_ARCH_LIST="3.5 5.2 6.0 6.1 7.0+PTX" TORCH_NVCC_FLAGS="-Xfatbin -compress-all" \
+RUN cd /opt/pytorch && \
+    TORCH_CUDA_ARCH_LIST="3.5 5.2 6.0 6.1 7.0+PTX" TORCH_NVCC_FLAGS="-Xfatbin -compress-all" \
     CMAKE_PREFIX_PATH="$(dirname $(which conda))/../" \
     pip install -v .
 RUN git clone https://github.com/pytorch/vision.git && cd vision && pip install -v .

I hope this is useful if this problem is reproducible for other people.

[vsoch]

This is super helpful! Would you mind doing a PR to change the file proper? Thank you!

[vsoch]

Thanks again for this fix @dbdean ! Closing the issue here too :)