NVIDIA-Merlin · albert17 · Mar 17, 2022 · Mar 9, 2022 · Mar 9, 2022 · Mar 9, 2022
diff --git a/ci/dockerfile.ci b/ci/dockerfile.ci
@@ -4,17 +4,17 @@ ARG IMAGE=nvcr.io/nvidia/tritonserver:${TRITON_VERSION}-py3
 FROM ${IMAGE}
 
 # Args
-ARG CORE_VER=main
-ARG RMM_VER=v21.12.00
 ARG CUDF_VER=v21.12.02
-ARG NVTAB_VER=main
-ARG NVTAB_BACKEND_VER=main
-ARG MODELS_VER=main
+ARG HWLOC_VER=2.4.1
+ARG RMM_VER=v21.12.00
+ARG CORE_VER=main
 ARG HUGECTR_VER=master
 ARG HUGECTR_BACKEND_VER=main
+ARG MODELS_VER=main
+ARG NVTAB_VER=main
+ARG NVTAB_BACKEND_VER=main
+ARG SYSTEMS_VER=main
 ARG TF4REC_VER=main
-ARG HWLOC_VER=2.4.1
-
 
 # Envs
 ENV CUDA_SHORT_VERSION=11.6
@@ -153,6 +153,11 @@ RUN git clone https://github.com/NVIDIA-Merlin/core.git /core/ && \
     cd /core/ && git checkout ${CORE_VER} && pip install -e . --no-deps
 ENV PYTHONPATH=/core:$PYTHONPATH
 
+# Install Merlin Systems
+RUN git clone https://github.com/NVIDIA-Merlin/systems.git /systems/ && \
+    cd /systems/ && git checkout ${SYSTEMS_VER} && pip install --no-deps -e .
+    ENV PYTHONPATH=/systems:$PYTHONPATH
+
 # Install NVTabular
 ENV PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION='python'
 RUN git clone https://github.com/NVIDIA-Merlin/NVTabular.git /nvtabular/ && \

diff --git a/ci/test_container.sh b/ci/test_container.sh
@@ -7,39 +7,23 @@ devices=$2
 # Unit tests #
 ##############
 
-## Test NVTabular
-pytest /nvtabular/tests/unit
+## Test Core
+/core/ci/test_unit.sh $container $devices
 
-## Test HugeCTR
-### Training container
-if [ "$container" == "merlin-training" ]; then
-    # layers_test && \ Running oom in blossom
-    checker_test && \
-    # data_reader_test && \ Need Multi-GPU
-    device_map_test && \
-    loss_test && \
-    optimizer_test && \
-    regularizers_test # && \
-    # parser_test && \ Needs Multi-GPU
-    # auc_test Needs Multi-GPU
-### Inference container
-# elif [ "$container" == "merlin-inference" ]; then
-    # HugeCTR - Deactivated until it is self-contained and it runs
-    # inference_test
-fi
+## Test NVTabular
+/nvtabular/ci/test_unit.sh $container $devices
 
 ## Test Transformers4Rec
-if [ "$container" != "merlin-training" ]; then
-    /transformers4rec/ci/test_unit.sh $container $devices
-fi
+/transformers4rec/ci/test_unit.sh $container $devices
 
 ## Test Models
-if [ "$container" != "merlin-training" ]; then
-    pip install coverage
-    chmod +x /models/ci/test_unit.sh
-    /models/ci/test_unit.sh $container $devices
-fi
+pip install coverage
+/models/ci/test_unit.sh $container $devices
 
+## Test HugeCTR
+if [ "$container" == "merlin-training" ]; then
+    /hugectr/ci/test_unit.sh $container $devices
+fi
 
 #####################
 # Integration tests #
@@ -48,13 +32,9 @@ fi
 # Test NVTabular 
 ## Not shared storage in blossom yet
 regex="merlin(.)*-inference"
-if [[ "$container" =~ $regex ]]; then
+if [[ ! "$container" =~ $regex ]]; then
     /nvtabular/ci/test_integration.sh $container $devices --report 1
 fi
-# Test HugeCTR
-# Waiting to sync integration tests with them
 
 # Test Transformers4Rec
-if [ "$container" != "merlin-training" ]; then
-    /transformers4rec/ci/test_integration.sh $container $devices
-fi
+/transformers4rec/ci/test_integration.sh $container $devices
diff --git a/docker/inference/dockerfile.ctr b/docker/inference/dockerfile.ctr
@@ -1,19 +1,23 @@
 # syntax=docker/dockerfile:1.2
 ARG TRITON_VERSION=22.02
-ARG IMAGE=nvcr.io/nvidia/tritonserver:${TRITON_VERSION}-py3-min
-FROM ${IMAGE}
+ARG FULL_IMAGE=nvcr.io/nvidia/tritonserver:${TRITON_VERSION}-py3
+ARG BASE_IMAGE=${FULL_IMAGE}-min
+FROM ${FULL_IMAGE} as full
+FROM ${BASE_IMAGE} as base
+
 
 # Args
-ARG CORE_VER=main
-ARG RMM_VER=v21.12.00
 ARG CUDF_VER=v21.12.02
+ARG RMM_VER=v21.12.00
+ARG CORE_VER=main
+ARG HUGECTR_VER=master
+ARG HUGECTR_BACKEND_VER=main
 ARG NVTAB_VER=main
 ARG NVTAB_BACKEND_VER=main
 ARG MODELS_VER=main
-ARG HUGECTR_VER=master
-ARG HUGECTR_BACKEND_VER=main
+ARG SYSTEMS_VER=main
 ARG TF4REC_VER=main
-#
+
 # Envs
 ENV LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64:/usr/local/lib:/repos/dist/lib
 ENV DEBIAN_FRONTEND=noninteractive
@@ -23,33 +27,61 @@ ENV CUDA_CUDA_LIBRARY=${CUDA_HOME}/lib64/stubs
 ENV PATH=${CUDA_HOME}/lib64/:${PATH}:${CUDA_HOME}/bin
 ENV PYTHONPATH=/usr/lib/python3.8/site-packages:$PYTHONPATH
 
-# Install packages
+# Install system packages
 ENV DEBIAN_FRONTEND=noninteractive
+
 RUN apt update -y --fix-missing && \
     apt install -y --no-install-recommends software-properties-common && \
-    apt-get install -y --no-install-recommends \
+    wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/cuda-ubuntu2004.pin && \
+    mv cuda-ubuntu2004.pin /etc/apt/preferences.d/cuda-repository-pin-600 && \
+    apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/7fa2af80.pub && \
+    add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/ /"
+
+RUN apt update -y --fix-missing && \
+    apt install -y --no-install-recommends \
         clang-format \
+        datacenter-gpu-manager \
+        libb64-dev \
         libboost-serialization-dev \
         libcurl4-openssl-dev \
+        libre2-dev \
         libssl-dev \
         libtbb-dev \
+        openssl \
+        pkg-config \
         protobuf-compiler \
         python3-dev \
         python3-pip \
-        rapidjson-dev &&\
-    apt-get autoremove -y && \
-    apt-get clean && \
+        rapidjson-dev && \
+    apt autoremove -y && \
+    apt clean && \
     rm -rf /var/lib/apt/lists/*
 
 RUN ln -s /usr/bin/python3 /usr/bin/python
 
 # Install multiple packages
+RUN pip install numba==0.55.1 numpy==1.21.5 --no-deps
 RUN pip install cupy-cuda115 nvidia-pyindex pybind11 pytest protobuf transformers==4.12 tensorflow-metadata 
-RUN pip install betterproto cachetools graphviz nvtx scipy sklearn
-RUN pip install pandas numba==0.55.1 numpy==1.21.5
+RUN pip install betterproto cachetools graphviz nvtx scipy sklearn pandas
 RUN pip install tritonclient[all] grpcio-channelz
 RUN pip install dask==2021.11.2 distributed==2021.11.2 dask[dataframe]==2021.11.2 dask-cuda
 RUN pip install git+https://github.com/rapidsai/asvdb.git@main
+RUN pip install numba==0.55.1 numpy==1.21.5 --no-deps
+
+
+# Triton Server
+WORKDIR /opt/tritonserver
+COPY --chown=1000:1000 --from=full /opt/tritonserver/LICENSE .
+COPY --chown=1000:1000 --from=full /opt/tritonserver/TRITON_VERSION .
+COPY --chown=1000:1000 --from=full /opt/tritonserver/NVIDIA_Deep_Learning_Container_License.pdf .
+COPY --chown=1000:1000 --from=full /opt/tritonserver/bin bin/
+COPY --chown=1000:1000 --from=full /opt/tritonserver/lib lib/
+COPY --chown=1000:1000 --from=full /opt/tritonserver/include include/
+COPY --chown=1000:1000 --from=full /opt/tritonserver/repoagents/ repoagents/
+COPY --chown=1000:1000 --from=full /opt/tritonserver/backends/python backends/python
+COPY --chown=1000:1000 --from=full /usr/bin/serve /usr/bin/.
+ENV PATH=/opt/tritonserver/bin:${PATH}:
+ENV LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/tritonserver/lib
 
 # Install cmake
 RUN apt remove --purge cmake -y && wget http://www.cmake.org/files/v3.21/cmake-3.21.1.tar.gz && \
@@ -138,14 +170,18 @@ RUN git clone https://github.com/NVIDIA-Merlin/core.git /core/ && \
     cd /core/ && git checkout ${CORE_VER} && pip install -e . --no-deps
 ENV PYTHONPATH=/core:$PYTHONPATH
 
+# Install Merlin Systems
+RUN git clone https://github.com/NVIDIA-Merlin/systems.git /systems/ && \
+    cd /systems/ && git checkout ${SYSTEMS_VER} && pip install --no-deps -e .
+    ENV PYTHONPATH=/systems:$PYTHONPATH
+
 # Install NVTabular
 ENV PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION='python'
 RUN git clone https://github.com/NVIDIA-Merlin/NVTabular.git /nvtabular/ && \
     cd /nvtabular/ && git checkout ${NVTAB_VER} && pip install -e . --no-deps
 ENV PYTHONPATH=/nvtabular:$PYTHONPATH
 
 # Install Transformers4Rec
-RUN pip install transformers==4.12
 RUN git clone https://github.com/NVIDIA-Merlin/Transformers4Rec.git /transformers4rec && \
     cd /transformers4rec/ && git checkout ${TF4REC_VER} && pip install -e . --no-deps
 ENV PYTHONPATH=/transformers4rec:$PYTHONPATH

diff --git a/docker/inference/dockerfile.tf b/docker/inference/dockerfile.tf
@@ -1,19 +1,22 @@
 # syntax=docker/dockerfile:1.2
 ARG TRITON_VERSION=22.02
-ARG IMAGE=nvcr.io/nvidia/tritonserver:${TRITON_VERSION}-tf2-python-py3
-FROM ${IMAGE}
+ARG FULL_IMAGE=nvcr.io/nvidia/tritonserver:${TRITON_VERSION}-py3
+ARG BASE_IMAGE=nvcr.io/nvidia/tritonserver:${TRITON_VERSION}-tf2-python-py3
+FROM ${FULL_IMAGE} as full
+FROM ${BASE_IMAGE} as bas
 
 # Args
-ARG CORE_VER=main
-ARG RMM_VER=v21.12.00
 ARG CUDF_VER=v21.12.02
-ARG NVTAB_VER=main
-ARG NVTAB_BACKEND_VER=main
-ARG MODELS_VER=main
+ARG RMM_VER=v21.12.00
+ARG CORE_VER=main
 ARG HUGECTR_VER=master
 ARG HUGECTR_BACKEND_VER=main
+ARG MODELS_VER=main
+ARG NVTAB_VER=main
+ARG NVTAB_BACKEND_VER=main
+ARG SYSTEMS_VER=main
 ARG TF4REC_VER=main
-#
+
 # Envs
 ENV LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64:/usr/local/lib:/repos/dist/lib
 ENV DEBIAN_FRONTEND=noninteractive
@@ -26,20 +29,21 @@ ENV PYTHONPATH=/usr/lib/python3.8/site-packages:$PYTHONPATH
 # Install packages
 ENV DEBIAN_FRONTEND=noninteractive
 RUN apt update -y --fix-missing && \
-    apt install -y --no-install-recommends software-properties-common && \
-    apt-get install -y --no-install-recommends \
+    apt install -y --no-install-recommends \
         clang-format \
         libboost-serialization-dev \
 	libexpat1-dev \
 	libsasl2-2 \
         libssl-dev \
         libtbb-dev \
+	openssl \
 	policykit-1 \
         protobuf-compiler \
         rapidjson-dev \
+	software-properties-common \
         zlib1g-dev && \
-    apt-get autoremove -y && \
-    apt-get clean && \
+    apt autoremove -y && \
+    apt clean && \
     rm -rf /var/lib/apt/lists/*
 
 RUN ln -s /usr/bin/python3 /usr/bin/python
@@ -52,6 +56,10 @@ RUN pip install tritonclient[all] grpcio-channelz
 RUN pip install dask==2021.11.2 distributed==2021.11.2 dask[dataframe]==2021.11.2 dask-cuda
 RUN pip install git+https://github.com/rapidsai/asvdb.git@main
 
+# Triton Server
+WORKDIR /opt/tritonserver
+COPY --chown=1000:1000 --from=full /opt/tritonserver/backends/fil backends/fil/
+
 # Install cmake
 RUN apt remove --purge cmake -y && wget http://www.cmake.org/files/v3.21/cmake-3.21.1.tar.gz && \
     tar xf cmake-3.21.1.tar.gz && cd cmake-3.21.1 && ./configure && make && make install
@@ -139,14 +147,18 @@ RUN git clone https://github.com/NVIDIA-Merlin/core.git /core/ && \
     cd /core/ && git checkout ${CORE_VER} && pip install -e . --no-deps
 ENV PYTHONPATH=/core:$PYTHONPATH
 
+# Install Merlin Systems
+RUN git clone https://github.com/NVIDIA-Merlin/systems.git /systems/ && \
+    cd /systems/ && git checkout ${SYSTEMS_VER} && pip install --no-deps -e .
+    ENV PYTHONPATH=/systems:$PYTHONPATH
+
 # Install NVTabular
 ENV PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION='python'
 RUN git clone https://github.com/NVIDIA-Merlin/NVTabular.git /nvtabular/ && \
     cd /nvtabular/ && git checkout ${NVTAB_VER} && pip install -e . --no-deps
 ENV PYTHONPATH=/nvtabular:$PYTHONPATH
 
 # Install Transformers4Rec
-RUN pip install transformers==4.12
 RUN git clone https://github.com/NVIDIA-Merlin/Transformers4Rec.git /transformers4rec && \
     cd /transformers4rec/ && git checkout ${TF4REC_VER} && pip install -e . --no-deps
 ENV PYTHONPATH=/transformers4rec:$PYTHONPATH

diff --git a/docker/inference/dockerfile.torch b/docker/inference/dockerfile.torch
@@ -4,16 +4,17 @@ ARG IMAGE=nvcr.io/nvidia/tritonserver:${TRITON_VERSION}-pyt-python-py3
 FROM ${IMAGE}
 
 # Args
-ARG CORE_VER=main
-ARG RMM_VER=v21.12.00
 ARG CUDF_VER=v21.12.02
-ARG NVTAB_VER=main
-ARG NVTAB_BACKEND_VER=main
-ARG MODELS_VER=main
+ARG RMM_VER=v21.12.00
+ARG CORE_VER=main
 ARG HUGECTR_VER=master
 ARG HUGECTR_BACKEND_VER=main
+ARG MODELS_VER=main
+ARG NVTAB_VER=main
+ARG NVTAB_BACKEND_VER=main
+ARG SYSTEMS_VER=main
 ARG TF4REC_VER=main
-#
+
 # Envs
 ENV LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64:/usr/local/lib:/repos/dist/lib
 ENV DEBIAN_FRONTEND=noninteractive
@@ -26,20 +27,21 @@ ENV PYTHONPATH=/usr/lib/python3.8/site-packages:$PYTHONPATH
 # Install packages
 ENV DEBIAN_FRONTEND=noninteractive
 RUN apt update -y --fix-missing && \
-    apt install -y --no-install-recommends software-properties-common && \
-    apt-get install -y --no-install-recommends \
+    apt install -y --no-install-recommends \
         clang-format \
         libboost-serialization-dev \
         libexpat1-dev \
         libsasl2-2 \
         libssl-dev \
         libtbb-dev \
+        openssl \
         policykit-1 \
         protobuf-compiler \
         rapidjson-dev \
+        software-properties-common \
         zlib1g-dev && \
-    apt-get autoremove -y && \
-    apt-get clean && \
+    apt autoremove -y && \
+    apt clean && \
     rm -rf /var/lib/apt/lists/*
 
 RUN ln -s /usr/bin/python3 /usr/bin/python
@@ -139,14 +141,18 @@ RUN git clone https://github.com/NVIDIA-Merlin/core.git /core/ && \
     cd /core/ && git checkout ${CORE_VER} && pip install -e . --no-deps
 ENV PYTHONPATH=/core:$PYTHONPATH
 
+# Install Merlin Systems
+RUN git clone https://github.com/NVIDIA-Merlin/systems.git /systems/ && \
+    cd /systems/ && git checkout ${SYSTEMS_VER} && pip install --no-deps -e .
+    ENV PYTHONPATH=/systems:$PYTHONPATH
+
 # Install NVTabular
 ENV PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION='python'
 RUN git clone https://github.com/NVIDIA-Merlin/NVTabular.git /nvtabular/ && \
     cd /nvtabular/ && git checkout ${NVTAB_VER} && pip install -e . --no-deps
 ENV PYTHONPATH=/nvtabular:$PYTHONPATH
 
 # Install Transformers4Rec
-RUN pip install transformers==4.12
 RUN git clone https://github.com/NVIDIA-Merlin/Transformers4Rec.git /transformers4rec && \
     cd /transformers4rec/ && git checkout ${TF4REC_VER} && pip install -e . --no-deps
 ENV PYTHONPATH=/transformers4rec:$PYTHONPATH