new base image

Author: subpic

Dockerfile

@@ -1,72 +1,76 @@
-FROM nvidia/cuda:9.0-cudnn7-runtime-ubuntu16.04
-# Contributors: Syed Navaid (https://github.com/syedahmad)
-# Vlad Hosu (https://github.com/subpic)
+FROM nvcr.io/nvidia/tensorflow:20.09-tf2-py3
 
+ENV DEBIAN_FRONTEND noninteractive
 RUN apt-get update && apt-get install -y \
-wget \
-vim \
-bzip2 \
-graphviz \
-mc \
-htop \
-emacs-nox \
-rsync \
-less \
-curl \
-#software-properties-common \
-build-essential \
- libfreetype6-dev \
- libpng12-dev \
- openssh-client
-
-RUN curl -s https://packagecloud.io/install/repositories/github/git-lfs/script.deb.sh | bash
-
-#Downgrade CUDA, TF issue: https://github.com/tensorflow/tensorflow/issues/17566#issuecomment-372490062
-RUN apt-get install --allow-downgrades --allow-change-held-packages -y libcudnn7=7.0.5.15-1+cuda9.0 git-lfs
-
-#Install MINICONDA
-RUN wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh -O Miniconda.sh && \
-/bin/bash Miniconda.sh -b -p /opt/conda && \
-rm Miniconda.sh
+wget \
+vim \
+bzip2 \
+graphviz \
+mc \
+htop \
+emacs-nox \
+rsync \
+less \
+curl \
+build-essential \
+openssh-server \
+imagemagick
 
+# Install MINICONDA
+RUN wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh -O Miniconda.sh && \
+ /bin/bash Miniconda.sh -b -p /opt/conda && \
+ rm Miniconda.sh
 ENV PATH /opt/conda/bin:$PATH
+RUN conda update -y -n base -c defaults conda
 
-#Install ANACONDA Environment
-RUN conda create -y -n jupyter_env python=3.6 nb_conda_kernels py-xgboost-gpu \
-anaconda ipykernel tensorflow-gpu keras-gpu matplotlib numpy scipy pandas \
-Pillow pydot munch scikit-learn && \
-/opt/conda/envs/jupyter_env/bin/pip install jupyter-tensorboard gitless opencv-python && \
-/opt/conda/envs/jupyter_env/bin/python -m ipykernel.kernelspec
-
-RUN conda create -y -n jupyter_env27 python=2.7 nb_conda_kernels py-xgboost-gpu \
-anaconda ipykernel tensorflow-gpu keras-gpu matplotlib numpy scipy pandas \
-Pillow pydot munch scikit-learn && \
-/opt/conda/envs/jupyter_env27/bin/pip install jupyter-tensorboard gitless opencv-python && \
-/opt/conda/envs/jupyter_env27/bin/python -m ipykernel.kernelspec
+# CREATE ENV and INSTALL
+RUN conda create -y -n jupyter_env python nb_conda_kernels py-xgboost-gpu \
+ anaconda ipykernel tensorflow-gpu keras-gpu matplotlib numpy scipy pandas \
+ Pillow pydot munch scikit-learn future \
+ && /opt/conda/envs/jupyter_env/bin/pip install jupytext jupyter-tensorboard \
+ gitless opencv-python tensorflow-gpu==2.4.1 \
+ && /opt/conda/envs/jupyter_env/bin/python -m ipykernel.kernelspec
 
 # Install OpenCV
 RUN apt-get update && apt-get install -y libopencv-dev && \
  echo 'ln /dev/null /dev/raw1394' >> ~/.bashrc
 
+RUN curl -sSL https://raw.githubusercontent.com/python-poetry/poetry/master/get-poetry.py | python -
+ENV PATH $HOME/.poetry/bin:$PATH
+
 RUN set -x \
  && cd /tmp \
  && wget https://github.com/sdg-mit/gitless/releases/download/v0.8.8/gl-v0.8.8-linux-x86_64.tar.gz \
  && tar -C /usr/bin/ -xvf gl-v0.8.8-linux-x86_64.tar.gz --strip-components=1 \
  && rm -f gl-v0.8.8-linux-x86_64.tar.gz
 
 COPY docker_deps/entrypoint.sh /root/
+COPY tests/test_libraries.py /tmp/
 COPY configs/jupyter_notebook_config.py /root/.jupyter/
 
-RUN chmod +x /root/entrypoint.sh
+RUN chmod +x /root/entrypoint.sh && \
+chown -R 777 /tmp/test_libraries.py
 
-ENV LC_ALL=C.UTF-8 LANG=C.UTF-8
-ENV PYTHONPATH='/mnt/home/research/:/mnt/home/:$PYTHONPATH'
-ENV PASSWORD unknownknowns
-ENV LD_LIBRARY_PATH /usr/local/cuda/lib64:$LD_LIBRARY_PATH
+RUN mkdir /var/run/sshd \
+ && echo 'root:jupyter-data-science' | chpasswd \
+ && sed -i 's/PermitRootLogin prohibit-password/PermitRootLogin no/' /etc/ssh/sshd_config \
+ && sed 's@session\s*required\s*pam_loginuid.so@session optional pam_loginuid.so@g' -i /etc/pam.d/sshd \
+ && echo "export VISIBLE=now" >> /etc/profile
+ENV NOTVISIBLE "in users profile"
+
+RUN useradd -rm -d /home/science -s /bin/bash -g root -G sudo -u 1000 science \
+ && echo 'science:jupyter-data-science' | chpasswd
+
+ENV LC_ALL=C.UTF-8 LANG=C.UTF-8 \
+ PYTHONPATH='/mnt/home/research/:/mnt/home/:$PYTHONPATH' \
+ PASSWORD=jupyter-data-science \
+ LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH 
 
 WORKDIR /mnt/home/
 
-# IPython
-EXPOSE 8888 6006
+# Jupyter + Tensorboard + SSHD
+EXPOSE 8888 6006 22
 
 CMD bash /root/entrypoint.sh
+
+

LICENSE

@@ -1,6 +1,6 @@
 MIT License
 
-Copyright (c) 2019 Vlad Hosu
+Copyright (c) 2019-2021 Vlad Hosu
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

README.md

@@ -1,13 +1,17 @@
-# Jupyter-Lab Data Science Dockerfile
-This is a jupyter-lab based setup for data science, including Anaconda environments for Python (v2 and v3), with Keras + Tensorflow GPU and XGBoost GPU.
+## Build
 
-After building the image, run using:
 ```
-nvidia-docker run -v $PWD:/mnt/home/ -p 8888:8888 -p 6006:6006 jupyter-data-science "$@"
+docker build -f Dockerfile -t jupyter-dl .
 ```
 
-The entrypoint script accepts Jupyter-Lab arguments:
+## Execute
 
 ```
-e.g. nvidia-docker run -v $PWD:/mnt/home/ -p 8888:8888 -p 6006:6006 jupyter-data-science --notebook-dir=/mnt/home
+nvidia-docker run -v $PWD:/mnt/home/ -p 8888:8888 -p 6006:6006 jupyter-dl
+```
+
+Note: The entrypoint script accepts Jupyter-Lab arguments
+
+```
+e.g. nvidia-docker run -v $PWD:/mnt/home/ -p 8888:8888 -p 6006:6006 jupyter-dl --notebook-dir=/mnt/home
 ```

build.sh

@@ -1 +1 @@
-docker build -f Dockerfile -t jupyter-data-science .
+docker build -f Dockerfile -t jupyter-dl .

docker_deps/entrypoint.sh

@@ -15,5 +15,6 @@
 # ==============================================================================
 
 conda init bash
-nohup /opt/conda/envs/jupyter_env/bin/tensorboard --logdir=/mnt/home/research/data/logs&
+nohup /opt/conda/envs/jupyter_env/bin/tensorboard --logdir=/mnt/home/logs&
+nohup /usr/sbin/sshd -D&
 /opt/conda/envs/jupyter_env/bin/jupyter-lab "$@" --allow-root

test.sh

@@ -0,0 +1 @@
+nvidia-docker run -v $PWD:/mnt/home/ -p 8888:8888 -p 6006:6006 jupyter-dl "$@"

tests/test_libraries.py

@@ -0,0 +1,32 @@
+import xgboost as xgb
+import sklearn
+import tensorflow
+import keras
+import numpy as np
+import unittest
+
+from tensorflow.python.client import device_lib
+from keras import backend as K
+
+
+def get_available_gpus():
+local_device_protos = device_lib.list_local_devices()
+return [x.name for x in local_device_protos if x.device_type == 'GPU']
+
+
+class TestGPULibraries(unittest.TestCase):
+
+def test_tensorflow(self):
+gpus = get_available_gpus()
+assert(len(gpus))
+
+def test_keras_gpu(self):
+gpus = K.tensorflow_backend._get_available_gpus()
+assert(len(gpus))
+
+def test_xgboost_gpu(self):
+gbm = xgb.XGBRegressor(silent=False, tree_method = 'gpu_hist')
+X = np.random.rand(10,10)
+y = np.random.rand(10)
+r = gbm.fit(X, y, verbose=True)
+assert(np.mean(np.abs(r.predict(X)-y)))