From f623b17232ef4179fbd8507df4538ad9f255132f Mon Sep 17 00:00:00 2001 From: Byungheon Jeong <byungheon.jeong@gmail.com> Date: Wed, 3 Jun 2020 16:53:55 -0700 Subject: [PATCH] initial --- .gitlab-ci.yml | 14 ++++++++++++ Dockerfile.dockerfile | 50 +++++++++++++++++++++++++++++++++++++++++++ mxnetDeployment.yaml | 46 +++++++++++++++++++++++++++++++++++++++ 3 files changed, 110 insertions(+) create mode 100755 .gitlab-ci.yml create mode 100755 Dockerfile.dockerfile create mode 100755 mxnetDeployment.yaml diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml new file mode 100755 index 0000000..bd58638 --- /dev/null +++ b/.gitlab-ci.yml @@ -0,0 +1,14 @@ +image: gcr.io/kaniko-project/executor:debug-v0.16.0 + +stages: + - build-and-push + +build-and-push-job: + stage: build-and-push + except: + changes: + - "**/*.yaml" + - "README.md" + script: + - echo "{\"auths\":{\"$CI_REGISTRY\":{\"username\":\"$CI_REGISTRY_USER\",\"password\":\"$CI_REGISTRY_PASSWORD\"}}}" > /kaniko/.docker/config.json + - /kaniko/executor --context $CI_PROJECT_DIR --dockerfile $CI_PROJECT_DIR/Dockerfile --destination $CI_REGISTRY_IMAGE:${CI_COMMIT_SHA:0:8} --destination $CI_REGISTRY_IMAGE:latest diff --git a/Dockerfile.dockerfile b/Dockerfile.dockerfile new file mode 100755 index 0000000..2416acd --- /dev/null +++ b/Dockerfile.dockerfile @@ -0,0 +1,50 @@ +FROM mxnet/python + +WORKDIR /userdata/kerasData/mxnet + +# Install system packages +RUN apt-get update && apt-get install -y --no-install-recommends \ + bzip2 \ + g++ \ + git \ + graphviz \ + libgl1-mesa-glx \ + libhdf5-dev \ + sudo\ + openmpi-bin \ + xvfb \ + screen \ + wget && \ + rm -rf /var/lib/apt/lists/* + + +# Install conda +ENV CONDA_DIR /opt/conda +ENV PATH $CONDA_DIR/bin:$PATH + +RUN wget --quiet --no-check-certificate https://repo.continuum.io/miniconda/Miniconda3-4.2.12-Linux-x86_64.sh && \ + echo "c59b3dd3cad550ac7596e0d599b91e75d88826db132e4146030ef471bb434e9a *Miniconda3-4.2.12-Linux-x86_64.sh" | sha256sum -c - && \ + /bin/bash /Miniconda3-4.2.12-Linux-x86_64.sh -f -b -p $CONDA_DIR && \ + rm Miniconda3-4.2.12-Linux-x86_64.sh && \ + echo export PATH=$CONDA_DIR/bin:'$PATH' > /etc/profile.d/conda.sh + +ARG python_version=3.5.2 + +RUN conda config --append channels conda-forge + +RUN conda install \ + bcolz \ + h5py \ + matplotlib \ + mkl \ + nose \ + notebook \ + Pillow \ + pandas \ + pydot \ + pygpu \ + pyyaml \ + scikit-learn \ + six \ + theano \ + mkdocs \ diff --git a/mxnetDeployment.yaml b/mxnetDeployment.yaml new file mode 100755 index 0000000..6641aee --- /dev/null +++ b/mxnetDeployment.yaml @@ -0,0 +1,46 @@ +apiVersion: apps/v1 +kind: Deployment +metadata: + name: MXNET-gpu + namespace: digits +spec: + replicas: 1 + selector: + matchLabels: + k8s-app: MXNET-gpu + template: + metadata: + labels: + k8s-app: MXNET-gpu + spec: + tolerations: + - key: "region" + operator: "Equal" + value: "allow" + effect: "NoSchedule" + containers: + - name: MXNET-gpu + image: gitlab-registry.nautilus.optiputer.net/ar-noc/MXNET:latest + securityContext: + capabilities: + add: + - NET_ADMIN + command: ["/bin/sh", "-c"] + args: ["sleep infinity"] + resources: + limits: + nvidia.com/gpu: 1 + memory: "64Gi" + cpu: "8" + requests: + memory: "10Gi" + cpu: "2" + + volumeMounts: + - mountPath: /userdata/kerasData + name: modeltraining + volumes: + - name: modeltraining + persistentVolumeClaim: + claimName: modeltraining + -- GitLab