Update README and versions for 22.04 branch

Dockerfile.sdk

@@ -29,7 +29,7 @@
 #
 
 # Base image on the minimum Triton container
-ARG BASE_IMAGE=nvcr.io/nvidia/tritonserver:22.03-py3-min
+ARG BASE_IMAGE=nvcr.io/nvidia/tritonserver:22.04-py3-min
 
 ARG TRITON_CLIENT_REPO_SUBDIR=clientrepo
 ARG TRITON_COMMON_REPO_TAG=main

TRITON_VERSION

@@ -1 +1 @@
-2.21.0dev
+2.21.0

build.py

@@ -86,9 +86,9 @@
 # Note: Not all sha ids would successfuly compile and work.
 #
 TRITON_VERSION_MAP = {
-    '2.21.0dev': (
-        '22.04dev',  # triton container
-        '22.03',  # upstream container
+    '2.21.0': (
+        '22.04',  # triton container
+        '22.04',  # upstream container
         '1.10.0',  # ORT
         '2021.4.582',  # ORT OpenVINO
         (('2021.4', None), ('2021.4', '2021.4.582'),
@@ -1732,7 +1732,7 @@ def get_tagged_backend(be, version):
         action='append',
         required=False,
         help=
-        'Include specified backend in build as <backend-name>[:<repo-tag>]. If <repo-tag> starts with "pull/" then it refers to a pull-request reference, otherwise <repo-tag> indicates the git tag/branch to use for the build. If the version is non-development then the default <repo-tag> is the release branch matching the container version (e.g. version 22.03 -> branch r22.03); otherwise the default <repo-tag> is "main" (e.g. version 22.03dev -> branch main).'
+        'Include specified backend in build as <backend-name>[:<repo-tag>]. If <repo-tag> starts with "pull/" then it refers to a pull-request reference, otherwise <repo-tag> indicates the git tag/branch to use for the build. If the version is non-development then the default <repo-tag> is the release branch matching the container version (e.g. version 22.04 -> branch r22.04); otherwise the default <repo-tag> is "main" (e.g. version 22.04dev -> branch main).'
     )
     parser.add_argument(
         '--build-multiple-openvino',
@@ -1746,14 +1746,14 @@ def get_tagged_backend(be, version):
         action='append',
         required=False,
         help=
-        'The version of a component to use in the build as <component-name>:<repo-tag>. <component-name> can be "common", "core", "backend" or "thirdparty". If <repo-tag> starts with "pull/" then it refers to a pull-request reference, otherwise <repo-tag> indicates the git tag/branch. If the version is non-development then the default <repo-tag> is the release branch matching the container version (e.g. version 22.03 -> branch r22.03); otherwise the default <repo-tag> is "main" (e.g. version 22.03dev -> branch main).'
+        'The version of a component to use in the build as <component-name>:<repo-tag>. <component-name> can be "common", "core", "backend" or "thirdparty". If <repo-tag> starts with "pull/" then it refers to a pull-request reference, otherwise <repo-tag> indicates the git tag/branch. If the version is non-development then the default <repo-tag> is the release branch matching the container version (e.g. version 22.04 -> branch r22.04); otherwise the default <repo-tag> is "main" (e.g. version 22.04dev -> branch main).'
     )
     parser.add_argument(
         '--repoagent',
         action='append',
         required=False,
         help=
-        'Include specified repo agent in build as <repoagent-name>[:<repo-tag>]. If <repo-tag> starts with "pull/" then it refers to a pull-request reference, otherwise <repo-tag> indicates the git tag/branch to use for the build. If the version is non-development then the default <repo-tag> is the release branch matching the container version (e.g. version 22.03 -> branch r22.03); otherwise the default <repo-tag> is "main" (e.g. version 22.03dev -> branch main).'
+        'Include specified repo agent in build as <repoagent-name>[:<repo-tag>]. If <repo-tag> starts with "pull/" then it refers to a pull-request reference, otherwise <repo-tag> indicates the git tag/branch to use for the build. If the version is non-development then the default <repo-tag> is the release branch matching the container version (e.g. version 22.04 -> branch r22.04); otherwise the default <repo-tag> is "main" (e.g. version 22.04dev -> branch main).'
     )
     parser.add_argument(
         '--no-force-clone',

deploy/aws/values.yaml

@@ -27,7 +27,7 @@
 replicaCount: 1
 
 image:
-  imageName: nvcr.io/nvidia/tritonserver:22.03-py3
+  imageName: nvcr.io/nvidia/tritonserver:22.04-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: "2.20.0"
+appVersion: "2.21.0"
 description: Triton Inference Server (Fleet Command)
 name: triton-inference-server
 # version is the Chart version; update when changing anything in the chart (semver)

deploy/fleetcommand/values.yaml

@@ -27,7 +27,7 @@
 replicaCount: 1
 
 image:
-  imageName: nvcr.io/nvidia/tritonserver:22.03-py3
+  imageName: nvcr.io/nvidia/tritonserver:22.04-py3
   pullPolicy: IfNotPresent
   numGpus: 1
   serverCommand: tritonserver
@@ -46,13 +46,13 @@ image:
     # Model Control Mode (Optional, default: none)
     #
     # To set model control mode, uncomment and configure below
-    # See https://github.com/triton-inference-server/server/blob/r22.03/docs/model_management.md
+    # See https://github.com/triton-inference-server/server/blob/r22.04/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/r22.03/README.md
+    # see https://github.com/triton-inference-server/server/blob/r22.04/README.md
     #  for more details
 
 service:

deploy/gcp/values.yaml

@@ -27,7 +27,7 @@
 replicaCount: 1
 
 image:
-  imageName: nvcr.io/nvidia/tritonserver:22.02-py3
+  imageName: nvcr.io/nvidia/tritonserver:22.04-py3
   pullPolicy: IfNotPresent
   modelRepositoryPath: gs://triton-inference-server-repository/model_repository
   numGpus: 1

deploy/gke-marketplace-app/benchmark/perf-analyzer-script/triton_client.yaml

@@ -33,7 +33,7 @@ metadata:
   namespace: default
 spec:
   containers:
-  - image: nvcr.io/nvidia/tritonserver:22.03-py3-sdk
+  - image: nvcr.io/nvidia/tritonserver:22.04-py3-sdk
     imagePullPolicy: Always
     name: nv-triton-client
     securityContext:

deploy/gke-marketplace-app/server-deployer/build_and_push.sh

@@ -26,9 +26,9 @@
 
 export REGISTRY=gcr.io/$(gcloud config get-value project | tr ':' '/')
 export APP_NAME=tritonserver
-export MAJOR_VERSION=2.20
-export MINOR_VERSION=2.20.0
-export NGC_VERSION=22.03-py3
+export MAJOR_VERSION=2.21
+export MINOR_VERSION=2.21.0
+export NGC_VERSION=22.04-py3
 
 docker pull nvcr.io/nvidia/$APP_NAME:$NGC_VERSION
 

deploy/gke-marketplace-app/server-deployer/chart/triton/Chart.yaml

@@ -25,7 +25,7 @@
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 apiVersion: v1
-appVersion: "2.20"
+appVersion: "2.21"
 description: Triton Inference Server
 name: triton-inference-server
-version: 2.20.0
+version: 2.21.0

deploy/gke-marketplace-app/server-deployer/chart/triton/values.yaml

@@ -31,14 +31,14 @@ maxReplicaCount: 3
 tritonProtocol: HTTP
 # HPA GPU utilization autoscaling target
 HPATargetAverageValue: 85
-modelRepositoryPath: gs://triton_sample_models/22_03
-publishedVersion: '2.20.0'
+modelRepositoryPath: gs://triton_sample_models/22_04
+publishedVersion: '2.21.0'
 gcpMarketplace: true
 
 image:
   registry: gcr.io
   repository: nvidia-ngc-public/tritonserver
-  tag: 22.03-py3
+  tag: 22.04-py3
   pullPolicy: IfNotPresent
   # modify the model repository here to match your GCP storage bucket
   numGpus: 1

deploy/gke-marketplace-app/server-deployer/data-test/schema.yaml

@@ -27,7 +27,7 @@
 x-google-marketplace:
   schemaVersion: v2
   applicationApiVersion: v1beta1
-  publishedVersion: '2.20.0'
+  publishedVersion: '2.21.0'
   publishedVersionMetadata:
     releaseNote: >-
       Initial release.
@@ -89,7 +89,7 @@ properties:
   modelRepositoryPath:
     type: string
     title: Bucket where models are stored. Please make sure the user/service account to create the GKE app has permission to this GCS bucket. Read Triton documentation on configs and formatting details, supporting TensorRT, TensorFlow, Pytorch, Onnx ... etc.
-    default: gs://triton_sample_models/22_03
+    default: gs://triton_sample_models/22_04
   image.ldPreloadPath:
     type: string
     title: Leave this empty by default. Triton allows users to create custom layers for backend such as TensorRT plugin or Tensorflow custom ops, the compiled shared library must be provided via LD_PRELOAD environment variable.

deploy/gke-marketplace-app/server-deployer/schema.yaml

@@ -89,7 +89,7 @@ properties:
   modelRepositoryPath:
     type: string
     title: Bucket where models are stored. Please make sure the user/service account to create the GKE app has permission to this GCS bucket. Read Triton documentation on configs and formatting details, supporting TensorRT, TensorFlow, Pytorch, Onnx ... etc.
-    default: gs://triton_sample_models/22_03
+    default: gs://triton_sample_models/22_04
   image.ldPreloadPath:
     type: string
     title: Leave this empty by default. Triton allows users to create custom layers for backend such as TensorRT plugin or Tensorflow custom ops, the compiled shared library must be provided via LD_PRELOAD environment variable.

docs/build.md

@@ -109,8 +109,8 @@ invocation builds all features, backends, and repository agents.
 If you are building on *main* branch then `<container tag>` will
 default to "main". If you are building on a release branch then
 `<container tag>` will default to the branch name. For example, if you
-are building on the r22.03 branch, `<container tag>` will default to
-r22.03. Therefore, you typically do not need to provide `<container
+are building on the r22.04 branch, `<container tag>` will default to
+r22.04. Therefore, you typically do not need to provide `<container
 tag>` at all (nor the preceding colon). You can use a different
 `<container tag>` for a component to instead use the corresponding
 branch/tag in the build. For example, if you have a branch called
@@ -282,8 +282,8 @@ python build.py --cmake-dir=<path/to/repo>/build --build-dir=/tmp/citritonbuild
 If you are building on *main* branch then '<container tag>' will
 default to "main". If you are building on a release branch then
 '<container tag>' will default to the branch name. For example, if you
-are building on the r22.03 branch, '<container tag>' will default to
-r22.03. Therefore, you typically do not need to provide '<container
+are building on the r22.04 branch, '<container tag>' will default to
+r22.04. Therefore, you typically do not need to provide '<container
 tag>' at all (nor the preceding colon). You can use a different
 '<container tag>' for a component to instead use the corresponding
 branch/tag in the build. For example, if you have a branch called

docs/compose.md

@@ -44,8 +44,8 @@ from source to get more exact customization.
 The `compose.py` script can be found in the [server repository](https://github.com/triton-inference-server/server).
 Simply clone the repository and run `compose.py` to create a custom container.
 Note: Created container version will depend on the branch that was cloned.
-For example branch [r22.03](https://github.com/triton-inference-server/server/tree/r22.03)
-should be used to create a image based on the NGC 22.03 Triton release.
+For example branch [r22.04](https://github.com/triton-inference-server/server/tree/r22.04)
+should be used to create a image based on the NGC 22.04 Triton release.
 
 `compose.py` provides `--backend`, `--repoagent` options that allow you to
 specify which backends and repository agents to include in the custom image.
@@ -62,7 +62,7 @@ will provide a container `tritonserver` locally. You can access the container wi
 $ docker run -it tritonserver:latest
 ```
 
-Note: If `compose.py` is run on release versions `r22.03` and earlier,
+Note: If `compose.py` is run on release versions `r22.04` and earlier,
 the resulting container will have DCGM version 2.2.3 installed.
 This may result in different GPU statistic reporting behavior.
 
@@ -76,19 +76,19 @@ For example, running
 ```
 python3 compose.py --backend tensorflow1 --repoagent checksum
 ```
-on branch [r22.03](https://github.com/triton-inference-server/server/tree/r22.03) pulls:
-- `min` container `nvcr.io/nvidia/tritonserver:22.03-py3-min`
-- `full` container `nvcr.io/nvidia/tritonserver:22.03-py3`
+on branch [r22.04](https://github.com/triton-inference-server/server/tree/r22.04) pulls:
+- `min` container `nvcr.io/nvidia/tritonserver:22.04-py3-min`
+- `full` container `nvcr.io/nvidia/tritonserver:22.04-py3`
 
 Alternatively, users can specify the version of Triton container to pull from any branch by either:
 1. Adding flag `--container-version <container version>` to branch
 ```
-python3 compose.py --backend tensorflow1 --repoagent checksum --container-version 22.03
+python3 compose.py --backend tensorflow1 --repoagent checksum --container-version 22.04
 ```
 2. Specifying `--image min,<min container image name> --image full,<full container image name>`.
    The user is responsible for specifying compatible `min` and `full` containers.
 ```
-python3 compose.py --backend tensorflow1 --repoagent checksum --image min,nvcr.io/nvidia/tritonserver:22.03-py3-min --image full,nvcr.io/nvidia/tritonserver:22.03-py3
+python3 compose.py --backend tensorflow1 --repoagent checksum --image min,nvcr.io/nvidia/tritonserver:22.04-py3-min --image full,nvcr.io/nvidia/tritonserver:22.04-py3
 ```
 Method 1 and 2 will result in the same composed container. Furthermore, `--image` flag overrides the `--container-version` flag when both are specified.
 

docs/custom_operations.md

@@ -64,7 +64,7 @@ simple way to ensure you are using the correct version of TensorRT is
 to use the [NGC TensorRT
 container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorrt)
 corresponding to the Triton container. For example, if you are using
-the 22.03 version of Triton, use the 22.03 version of the TensorRT
+the 22.04 version of Triton, use the 22.04 version of the TensorRT
 container.
 
 ## TensorFlow
@@ -108,7 +108,7 @@ simple way to ensure you are using the correct version of TensorFlow
 is to use the [NGC TensorFlow
 container](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow)
 corresponding to the Triton container. For example, if you are using
-the 22.03 version of Triton, use the 22.03 version of the TensorFlow
+the 22.04 version of Triton, use the 22.04 version of the TensorFlow
 container.
 
 ## PyTorch
@@ -152,7 +152,7 @@ simple way to ensure you are using the correct version of PyTorch is
 to use the [NGC PyTorch
 container](https://ngc.nvidia.com/catalog/containers/nvidia:pytorch)
 corresponding to the Triton container. For example, if you are using
-the 22.03 version of Triton, use the 22.03 version of the PyTorch
+the 22.04 version of Triton, use the 22.04 version of the PyTorch
 container.
 
 ## ONNX

docs/test.md

@@ -49,7 +49,7 @@ $ ./gen_qa_custom_ops
 ```
 
 This will create multiple model repositories in /tmp/<version>/qa_*
-(for example /tmp/22.03/qa_model_repository).  The TensorRT models
+(for example /tmp/22.04/qa_model_repository).  The TensorRT models
 will be created for the GPU on the system that CUDA considers device 0
 (zero). If you have multiple GPUs on your system see the documentation
 in the scripts for how to target a specific GPU.

qa/common/gen_qa_custom_ops

@@ -37,7 +37,7 @@
 ##
 ############################################################################
 
-TRITON_VERSION=22.03
+TRITON_VERSION=22.04
 TENSORFLOW_IMAGE=${TENSORFLOW_IMAGE:=nvcr.io/nvidia/tensorflow:$TRITON_VERSION-tf1-py3}
 PYTORCH_IMAGE=${PYTORCH_IMAGE:=nvcr.io/nvidia/pytorch:$TRITON_VERSION-py3}
 

qa/common/gen_qa_model_repository

@@ -48,7 +48,7 @@
 ##
 ############################################################################
 
-TRITON_VERSION=22.03
+TRITON_VERSION=22.04
 
 # ONNX. Use ONNX_OPSET 0 to use the default for ONNX version
 ONNX_VERSION=1.10.1