Docker: tensorflow gpu version, size reduction and documentation upda…

…tes (#18)

* Docker Documentation

* Docker image size

* Docker image size reducing

* Docker size

* bug in tensorflow GPU fixed

* 4Gb to download

Dockerfile

@@ -3,35 +3,39 @@ FROM nvidia/cuda:9.0-cudnn7-devel-ubuntu16.04
 ENV DEBIAN_FRONTEND noninteractive
 ENV TERM linux
 
-RUN mkdir /code
-COPY . /code
+RUN mkdir /moses
+COPY . /moses
 
 RUN set -ex \
     && apt-get update -yqq \
     && apt-get upgrade -yqq \
     && apt-get install -yqq --no-install-recommends \
-        git wget curl libxrender1 libxext6 software-properties-common \
+        git wget curl ssh libxrender1 libxext6 software-properties-common \
     && wget --no-check-certificate https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh \
     && /bin/bash Miniconda3-latest-Linux-x86_64.sh -f -b -p /opt/miniconda \
     && add-apt-repository ppa:git-core/ppa \
     && (curl -s https://packagecloud.io/install/repositories/github/git-lfs/script.deb.sh | bash) \
     && apt-get install git-lfs \
     && git lfs install \
     && apt-get clean \ 
-    && rm Miniconda3-latest-Linux-x86_64.sh
+    && /opt/miniconda/bin/conda install conda=4.5.12=py36_0 \
+    && /opt/miniconda/bin/conda clean -yq -a \
+    && rm Miniconda3-latest-Linux-x86_64.sh \ 
+    && rm -rf \
+        /tmp/* \
+        /var/tmp/* \
+        /usr/share/man \
+        /usr/share/doc \
+        /usr/share/doc-base
 
 ENV PATH /opt/miniconda/bin:$PATH
-RUN conda update conda
-
-RUN conda install -y -q numpy=1.15.0 scipy=1.1.0 matplotlib=3.0.1 pandas=0.23.3 scikit-learn=0.19.1 tqdm
-RUN conda install -c anaconda tensorflow-gpu=1.12
-RUN conda install -y -q keras-gpu=2.2.4
-RUN conda install -y -q -c rdkit rdkit=2018.09.1.0
-RUN conda install -y -q -c pytorch pytorch=0.4.1
-RUN conda install -y -q -c pytorch torchvision=0.2.1
-
-RUN cd /code && python setup.py install && git lfs pull
-
-RUN conda clean -yq -a
+RUN conda install -yq numpy=1.15.0 scipy=1.1.0 matplotlib=3.0.1 pandas=0.23.3 scikit-learn=0.19.1 tqdm \
+    && pip install tensorflow-gpu==1.12 keras==2.2.4 \
+    && conda install -yq -c rdkit rdkit=2018.09.1.0 \
+    && conda install -yq -c pytorch pytorch=0.4.1 torchvision=0.2.1 \
+    && conda clean -yq -a
+
+WORKDIR /moses
+RUN python setup.py install && git lfs pull && conda clean -yq -a && rm -rf .git/lfs
 
 CMD [ "/bin/bash" ]

README.md

@@ -150,27 +150,25 @@ For comparison of molecular properties, we computed the Frèchet distance betwee
 
 1. Install [docker](https://docs.docker.com/install/) and [nvidia-docker](https://github.com/nvidia/nvidia-docker/wiki/Installation-(version-2.0)).
 
-2. Pull an existing image from DockerHub:
+2. Pull an existing image (4.1Gb to download) from DockerHub:
 
 ```
 docker pull molecularsets/moses
 ```
 
 or clone the repository and build it manually:
 
-
 ```
-git lfs install
 git clone https://github.com/molecularsets/moses.git
 nvidia-docker image build --tag molecularsets/moses moses/
 ```
 
 3. Create a container:
 ```
-nvidia-docker run -it moses --network="host" --shm-size 1G molecularsets/moses
+nvidia-docker run -it --name moses --network="host" --shm-size 1G molecularsets/moses
 ```
 
-4. The dataset and source code is available inside the docker container:
+4. The dataset and source code are available inside the docker container at /moses:
 ```
 docker exec -it molecularsets/moses bash
 ```