remake nvidia docker (#6686)

* use latest

* remake

* examples

azure-pipelines.yml

@@ -98,8 +98,9 @@ jobs:
     - script: |
         set -e
         python -m pytest pl_examples -v --maxfail=2 --durations=0
-        python setup.py install --user --quiet
-        bash pl_examples/run_ddp-example.sh
+        pip install . --user --quiet
+        bash pl_examples/run_examples-args.sh --gpus 1 --max_epochs 1 --batch_size 64 --limit_train_batches 5 --limit_val_batches 3
+        bash pl_examples/run_ddp-examples.sh --max_epochs 1 --batch_size 32 --limit_train_batches 2 --limit_val_batches 2
         # cd pl_examples/basic_examples
         # bash submit_ddp_job.sh
         # bash submit_ddp2_job.sh

dockers/nvidia/Dockerfile

@@ -12,26 +12,69 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-FROM nvcr.io/nvidia/pytorch:20.12-py3
+FROM nvcr.io/nvidia/cuda:11.1.1-runtime-ubuntu20.04
 
 MAINTAINER PyTorchLightning <https://github.com/PyTorchLightning>
 
 ARG LIGHTNING_VERSION=""
 
-COPY ./ ./pytorch-lightning/
+SHELL ["/bin/bash", "-c"]
+# https://techoverflow.net/2019/05/18/how-to-fix-configuring-tzdata-interactive-input-when-building-docker-images/
+ENV \
+    DEBIAN_FRONTEND=noninteractive \
+    TZ=Europe/Prague \
+    PATH="$PATH:/root/.local/bin" \
+    CUDA_TOOLKIT_ROOT_DIR="/usr/local/cuda" \
+    MKL_THREADING_LAYER=GNU
+
+RUN apt-get update -qq && \
+    apt-get install -y --no-install-recommends \
+        build-essential \
+        python3 \
+        python3-distutils \
+        python3-dev \
+        pkg-config \
+        cmake \
+        git \
+        wget \
+        unzip \
+        ca-certificates \
+    && \
+
+# Cleaning
+    apt-get autoremove -y && \
+    apt-get clean && \
+    rm -rf /root/.cache && \
+    rm -rf /var/lib/apt/lists/* && \
+
+# Setup PIP
+    update-alternatives --install /usr/bin/python python /usr/bin/python3 1 && \
+    wget https://bootstrap.pypa.io/get-pip.py --progress=bar:force:noscroll --no-check-certificate && \
+    python get-pip.py && \
+    rm get-pip.py && \
+    pip --version
+
+COPY ./ /home/pytorch-lightning/
 
-# install dependencies
 RUN \
-    # Disable cache
-    #conda install "pip>20.1" && \
-    #pip config set global.cache-dir false && \
-    if [ -z $LIGHTNING_VERSION ] ; then \
-        pip install ./pytorch-lightning --no-cache-dir ; \
-        rm -rf pytorch-lightning ; \
-    else \
+    cd /home  && \
+    mv pytorch-lightning/notebooks . && \
+    mv pytorch-lightning/pl_examples . && \
+    # replace by specific version if asked
+    if [ ! -z "$LIGHTNING_VERSION" ] ; then \
         rm -rf pytorch-lightning ; \
-        pip install https://github.com/PyTorchLightning/pytorch-lightning/archive/${LIGHTNING_VERSION}.zip --no-cache-dir ; \
-    fi
+        wget https://github.com/PyTorchLightning/pytorch-lightning/archive/${LIGHTNING_VERSION}.zip --progress=bar:force:noscroll ; \
+        unzip ${LIGHTNING_VERSION}.zip ; \
+        mv pytorch-lightning-*/ pytorch-lightning ; \
+        rm *.zip ; \
+    fi && \
+
+# Installations
+    python -c "fname = './pytorch-lightning/requirements/extra.txt' ; lines = [line for line in open(fname).readlines() if not line.startswith('horovod')] ; open(fname, 'w').writelines(lines)" && \
+    pip install -r ./pytorch-lightning/requirements/extra.txt -U --no-cache-dir && \
+    pip install -r ./pytorch-lightning/requirements/examples.txt -U --no-cache-dir && \
+    pip install ./pytorch-lightning --no-cache-dir && \
+    rm -rf pytorch-lightning
 
 RUN python --version && \
     pip --version && \

dockers/release/Dockerfile

@@ -21,12 +21,14 @@ MAINTAINER PyTorchLightning <https://github.com/PyTorchLightning>
 
 ARG LIGHTNING_VERSION=""
 
-COPY ./ ./pytorch-lightning/
+COPY ./ /home/pytorch-lightning/
 
 # install dependencies
 RUN \
-    # Disable cache
-    #conda install "pip>20.1" && \
+    cd /home && \
+    mv pytorch-lightning/notebooks . && \
+    mv pytorch-lightning/pl_examples . && \
+    # replace by specific version if asked
     if [ ! -z "$LIGHTNING_VERSION" ] ; then \
         rm -rf pytorch-lightning ; \
         wget https://github.com/PyTorchLightning/pytorch-lightning/archive/${LIGHTNING_VERSION}.zip --progress=bar:force:noscroll ; \

pl_examples/run_ddp-examples.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+ARGS_EXTRA_DDP=" --gpus 2 --accelerator ddp"
+ARGS_EXTRA_AMP=" --precision 16"
+
+python pl_examples/basic_examples/simple_image_classifier.py $@ ${ARGS_EXTRA_DDP}
+python pl_examples/basic_examples/simple_image_classifier.py $@ ${ARGS_EXTRA_DDP} ${ARGS_EXTRA_AMP}
+
+python pl_examples/basic_examples/backbone_image_classifier.py $@ ${ARGS_EXTRA_DDP}
+python pl_examples/basic_examples/backbone_image_classifier.py $@ ${ARGS_EXTRA_DDP} ${ARGS_EXTRA_AMP}
+
+python pl_examples/basic_examples/autoencoder.py $@ ${ARGS_EXTRA_DDP}
+python pl_examples/basic_examples/autoencoder.py $@ ${ARGS_EXTRA_DDP} ${ARGS_EXTRA_AMP}

pl_examples/run_examples-args.sh

@@ -0,0 +1,15 @@
+#!/bin/bash
+
+echo $@
+
+full_path=$(realpath $0)
+echo $full_path
+
+dir_path=$(dirname $full_path)
+echo $dir_path
+
+python ${dir_path}/basic_examples/simple_image_classifier.py $@
+
+python ${dir_path}/basic_examples/backbone_image_classifier.py $@
+
+python ${dir_path}/basic_examples/autoencoder.py $@