Merge branch 'main' of github.com:triton-inference-server/server into yinggeh-DLIS-7272-refactor-core-input-checks

Author: yinggeh

Dockerfile.QA

@@ -390,6 +390,10 @@ RUN rm -fr qa/L0_copyrights qa/L0_build_variants && \
 RUN find qa/pkgs/ -maxdepth 1 -type f -name \
     "tritonserver-*.whl" | xargs -I {} pip3 install --upgrade {}[all]
 
+# Install Triton Frontend Python API
+RUN find qa/pkgs/ -type f -name \
+    "tritonfrontend-*.whl" | xargs -I {} pip3 install --upgrade {}[all]
+
 ENV LD_LIBRARY_PATH /opt/tritonserver/qa/clients:${LD_LIBRARY_PATH}
 
 # DLIS-3631: Needed to run Perf Analyzer CI tests correctly

Dockerfile.sdk

@@ -29,7 +29,7 @@
 #
 
 # Base image on the minimum Triton container
-ARG BASE_IMAGE=nvcr.io/nvidia/tritonserver:24.07-py3-min
+ARG BASE_IMAGE=nvcr.io/nvidia/tritonserver:24.08-py3-min
 
 ARG TRITON_CLIENT_REPO_SUBDIR=clientrepo
 ARG TRITON_PA_REPO_SUBDIR=perfanalyzerrepo

Dockerfile.win10.min

@@ -37,7 +37,7 @@ RUN choco install unzip -y
 #
 # Installing TensorRT
 #
-ARG TENSORRT_VERSION=10.2.0.19
+ARG TENSORRT_VERSION=10.3.0.26
 ARG TENSORRT_ZIP="TensorRT-${TENSORRT_VERSION}.Windows10.x86_64.cuda-12.5.zip"
 ARG TENSORRT_SOURCE=https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/10.2.0/zip/TensorRT-10.2.0.19.Windows10.x86_64.cuda-12.5.zip
 # COPY ${TENSORRT_ZIP} /tmp/${TENSORRT_ZIP}
@@ -51,7 +51,7 @@ LABEL TENSORRT_VERSION="${TENSORRT_VERSION}"
 #
 # Installing cuDNN
 #
-ARG CUDNN_VERSION=9.2.1.18
+ARG CUDNN_VERSION=9.3.0.75
 ARG CUDNN_ZIP=cudnn-windows-x86_64-${CUDNN_VERSION}_cuda12-archive.zip
 ARG CUDNN_SOURCE=https://developer.download.nvidia.com/compute/cudnn/redist/cudnn/windows-x86_64/cudnn-windows-x86_64-9.2.1.18_cuda12-archive.zip
 ADD ${CUDNN_SOURCE} /tmp/${CUDNN_ZIP}
@@ -101,14 +101,14 @@ LABEL CMAKE_VERSION=${CMAKE_VERSION}
 #
 # Installing Visual Studio BuildTools: VS17 2022
 #
-ARG BUILDTOOLS_VERSION=17.9.34622.214
+ARG BUILDTOOLS_VERSION=17.10.35201.131
 # Download collect.exe in case of an install failure.
 ADD https://aka.ms/vscollect.exe "C:\tmp\collect.exe"
 
 # Use the latest release channel. For more control, specify the location of an internal layout.
 # Download the Build Tools bootstrapper.
 # ARG BUILD_TOOLS_SOURCE=https://aka.ms/vs/17/release/vs_buildtools.exe
-ARG BUILD_TOOLS_SOURCE=https://download.visualstudio.microsoft.com/download/pr/5e7b923b-7d89-4e14-95b8-a84ab168e243/96b21d216c7954aaf606c6d7ba59a3de991884a8a86c578c767ba349c23188a9/vs_BuildTools.exe
+ARG BUILD_TOOLS_SOURCE=https://download.visualstudio.microsoft.com/download/pr/28626b4b-f88f-4b55-a0cf-f3eaa2c643fb/e6c43d4dfb36338d954cdb3ad9010ab2a479e712088f4f6b016eadcc721bab28/vs_BuildTools.exe
 ADD ${BUILD_TOOLS_SOURCE} vs_buildtools.exe
 # Install Build Tools with the Microsoft.VisualStudio.Workload.VCTools workload, including recommended.
 ARG VS_INSTALL_PATH_WP="C:\BuildTools"
@@ -175,15 +175,15 @@ RUN copy "%CUDA_INSTALL_ROOT_WP%\extras\visual_studio_integration\MSBuildExtensi
 
 RUN setx PATH "%CUDA_INSTALL_ROOT_WP%\bin;%PATH%"
 
-ARG CUDNN_VERSION=9.2.1.18
+ARG CUDNN_VERSION=9.3.0.75
 ENV CUDNN_VERSION ${CUDNN_VERSION}
 COPY --from=dependency_base /cudnn /cudnn
 RUN copy cudnn\bin\cudnn*.dll "%CUDA_INSTALL_ROOT_WP%\bin\."
 RUN copy cudnn\lib\x64\cudnn*.lib "%CUDA_INSTALL_ROOT_WP%\lib\x64\."
 RUN copy cudnn\include\cudnn*.h "%CUDA_INSTALL_ROOT_WP%\include\."
 LABEL CUDNN_VERSION="${CUDNN_VERSION}"
 
-ARG TENSORRT_VERSION=10.2.0.19
+ARG TENSORRT_VERSION=10.3.0.26
 ENV TRT_VERSION ${TENSORRT_VERSION}
 COPY --from=dependency_base /TensorRT /TensorRT
 RUN setx PATH "c:\TensorRT\lib;%PATH%"

README.md

@@ -45,7 +45,7 @@ ___
 
 ##### LATEST RELEASE
 You are currently on the `main` branch which tracks under-development progress towards the next release.
-The current release is version [2.48.0](https://github.com/triton-inference-server/server/releases/latest) and corresponds to the 24.07 container release on NVIDIA GPU Cloud (NGC).
+The current release is version [2.49.0](https://github.com/triton-inference-server/server/releases/latest) and corresponds to the 24.08 container release on NVIDIA GPU Cloud (NGC).
 
 Triton Inference Server is an open source inference serving software that
 streamlines AI inferencing. Triton enables teams to deploy any AI model from
@@ -103,16 +103,16 @@ Inference Server with the
 
 ```bash
 # Step 1: Create the example model repository
-git clone -b r24.07 https://github.com/triton-inference-server/server.git
+git clone -b r24.08 https://github.com/triton-inference-server/server.git
 cd server/docs/examples
 ./fetch_models.sh
 
 # Step 2: Launch triton from the NGC Triton container
-docker run --gpus=1 --rm --net=host -v ${PWD}/model_repository:/models nvcr.io/nvidia/tritonserver:24.07-py3 tritonserver --model-repository=/models
+docker run --gpus=1 --rm --net=host -v ${PWD}/model_repository:/models nvcr.io/nvidia/tritonserver:24.08-py3 tritonserver --model-repository=/models
 
 # Step 3: Sending an Inference Request
 # In a separate console, launch the image_client example from the NGC Triton SDK container
-docker run -it --rm --net=host nvcr.io/nvidia/tritonserver:24.07-py3-sdk
+docker run -it --rm --net=host nvcr.io/nvidia/tritonserver:24.08-py3-sdk
 /workspace/install/bin/image_client -m densenet_onnx -c 3 -s INCEPTION /workspace/images/mug.jpg
 
 # Inference should return the following

TRITON_VERSION

@@ -1 +1 @@
-2.49.0dev
+2.50.0dev

build.py

@@ -70,9 +70,9 @@
 # incorrectly load the other version of the openvino libraries.
 #
 TRITON_VERSION_MAP = {
-    "2.49.0dev": (
-        "24.08dev",  # triton container
-        "24.07",  # upstream container
+    "2.50.0dev": (
+        "24.09dev",  # triton container
+        "24.08",  # upstream container
         "1.18.1",  # ORT
         "2024.0.0",  # ORT OpenVINO
         "2024.0.0",  # Standalone OpenVINO
@@ -116,7 +116,8 @@ def fail_if(p, msg):
 
 
 def target_platform():
-    if FLAGS.target_platform is not None:
+    # When called by compose.py, FLAGS will be None
+    if FLAGS and FLAGS.target_platform is not None:
         return FLAGS.target_platform
     platform_string = platform.system().lower()
     if platform_string == "linux":
@@ -132,7 +133,8 @@ def target_platform():
 
 
 def target_machine():
-    if FLAGS.target_machine is not None:
+    # When called by compose.py, FLAGS will be None
+    if FLAGS and FLAGS.target_machine is not None:
         return FLAGS.target_machine
     return platform.machine().lower()
 
@@ -214,6 +216,8 @@ def header(self, desc=None):
 
         self.comment("Exit script immediately if any command fails")
         if target_platform() == "windows":
+            self._file.write("$UseStructuredOutput = $false\n")
+            self.blankln()
             self._file.write("function ExitWithCode($exitcode) {\n")
             self._file.write("    $host.SetShouldExit($exitcode)\n")
             self._file.write("    exit $exitcode\n")
@@ -639,13 +643,16 @@ def pytorch_cmake_args(images):
         cmake_backend_arg("pytorch", "TRITON_PYTORCH_DOCKER_IMAGE", None, image),
     ]
 
-    if FLAGS.enable_gpu:
+    # TODO: TPRD-372 TorchTRT extension is not currently supported by our manylinux build
+    # TODO: TPRD-373 NVTX extension is not currently supported by our manylinux build
+    if target_platform() != "rhel":
+        if FLAGS.enable_gpu:
+            cargs.append(
+                cmake_backend_enable("pytorch", "TRITON_PYTORCH_ENABLE_TORCHTRT", True)
+            )
         cargs.append(
-            cmake_backend_enable("pytorch", "TRITON_PYTORCH_ENABLE_TORCHTRT", True)
+            cmake_backend_enable("pytorch", "TRITON_ENABLE_NVTX", FLAGS.enable_nvtx)
         )
-    cargs.append(
-        cmake_backend_enable("pytorch", "TRITON_ENABLE_NVTX", FLAGS.enable_nvtx)
-    )
     return cargs
 
 
@@ -655,7 +662,9 @@ def onnxruntime_cmake_args(images, library_paths):
             "onnxruntime",
             "TRITON_BUILD_ONNXRUNTIME_VERSION",
             None,
-            TRITON_VERSION_MAP[FLAGS.version][2],
+            os.getenv("TRITON_BUILD_ONNXRUNTIME_VERSION")
+            if os.getenv("TRITON_BUILD_ONNXRUNTIME_VERSION")
+            else TRITON_VERSION_MAP[FLAGS.version][2],
         )
     ]
 
@@ -1301,7 +1310,6 @@ def dockerfile_prepare_container_linux(argmap, backends, enable_gpu, target_mach
         gpu_enabled=gpu_enabled
     )
 
-    # This
     if target_platform() == "rhel":
         df += """
 # Common dpeendencies.

deploy/aws/values.yaml

@@ -27,7 +27,7 @@
 replicaCount: 1
 
 image:
-  imageName: nvcr.io/nvidia/tritonserver:24.07-py3
+  imageName: nvcr.io/nvidia/tritonserver:24.08-py3
   pullPolicy: IfNotPresent
   modelRepositoryPath: s3://triton-inference-server-repository/model_repository
   numGpus: 1

deploy/fleetcommand/Chart.yaml

@@ -26,7 +26,7 @@
 
 apiVersion: v1
 # appVersion is the Triton version; update when changing release
-appVersion: "2.48.0"
+appVersion: "2.49.0"
 description: Triton Inference Server (Fleet Command)
 name: triton-inference-server
 # version is the Chart version; update when changing anything in the chart

deploy/fleetcommand/values.yaml

@@ -27,7 +27,7 @@
 replicaCount: 1
 
 image:
-  imageName: nvcr.io/nvidia/tritonserver:24.07-py3
+  imageName: nvcr.io/nvidia/tritonserver:24.08-py3
   pullPolicy: IfNotPresent
   numGpus: 1
   serverCommand: tritonserver
@@ -47,13 +47,13 @@ image:
     #
     # To set model control mode, uncomment and configure below
     # TODO: Fix the following url, it is invalid
-    # See https://github.com/triton-inference-server/server/blob/r24.07/docs/model_management.md
+    # See https://github.com/triton-inference-server/server/blob/r24.08/docs/model_management.md
     #  for more details
     #- --model-control-mode=explicit|poll|none
     #
     # Additional server args
     #
-    # see https://github.com/triton-inference-server/server/blob/r24.07/README.md
+    # see https://github.com/triton-inference-server/server/blob/r24.08/README.md
     #  for more details
 
 service:

deploy/gcp/values.yaml

@@ -27,7 +27,7 @@
 replicaCount: 1
 
 image:
-  imageName: nvcr.io/nvidia/tritonserver:24.07-py3
+  imageName: nvcr.io/nvidia/tritonserver:24.08-py3
   pullPolicy: IfNotPresent
   modelRepositoryPath: gs://triton-inference-server-repository/model_repository
   numGpus: 1