Install Instructions for Docker fail

🐛 Bug

The Docker instructions in the INSTALL.md file do not work. The following command was ran:

nvidia-docker build -t maskrcnn-benchmark --build-arg CUDA=9.2 --build-arg CUDNN=7 docker/

The following error was received on Step 8:

Step 8/23 : RUN /miniconda/bin/conda install -y conda-build && /miniconda/bin/conda create -y --name py36 python=3.6.7 && /miniconda/bin/conda clean -ya ---> Running in 67403eddebe3

/bin/sh: 1: /miniconda/bin/conda: not found

Expected behavior

A clean Installation

Environment

Collecting environment information… PyTorch version: N/A Is debug build: N/A CUDA used to build PyTorch: N/A

OS: CentOS Linux 7 (Core) GCC version: (GCC) 4.8.5 20150623 (Red Hat 4.8.5-11) CMake version: Could not collect

Python version: 2.7 Is CUDA available: N/A CUDA runtime version: Could not collect GPU models and configuration: GPU 0: Tesla K40m Nvidia driver version: 440.64.00 cuDNN version: Could not collect

Additional context

[midpushgogo]

在dockerfile中修改第14行即可解决问题： RUN curl -L -so /miniconda.sh https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh \

原因：原url需要http重定向才能下载，curl默认不进行重定向

[andics]

@sbkim052

Unfortunately, this repo is pretty dead - I don’t think the creators will be fixing much anytime soon… I did find a solution, however: https://ngc.nvidia.com/catalog/resources/nvidia:mask_r_cnn_for_pytorch/quickStartGuide https://github.com/NVIDIA/DeepLearningExamples/tree/master/PyTorch/Segmentation/MaskRCNN#setup

This is a distribution provided directly by NVIDIA. I managed to set it up with no problems. The source code is slightly different as it uses newer versions of PyTorch and other libraries. The architecture is the same though, so you can rest assured it is still MaskRCNN you’re using. Hope that helps!