Skip to content

Commit 1a34a63

Browse files
mc-nvmc-nv
committed
Update README for 22.07 release
1 parent 5fdd7b0 commit 1a34a63

File tree

2 files changed

+340
-2
lines changed

2 files changed

+340
-2
lines changed

‎README.md

+218-2
Original file line numberDiff line numberDiff line change
@@ -26,5 +26,221 @@
2626
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
2727
-->
2828

29-
**NOTE: You are currently on the r22.07 branch which tracks stabilization
30-
towards the next release. This branch is not usable during stabilization.**
29+
# Triton Inference Server
30+
31+
[![License](https://img.shields.io/badge/License-BSD3-lightgrey.svg)](https://opensource.org/licenses/BSD-3-Clause)
32+
33+
**LATEST RELEASE: You are currently on the main branch which tracks
34+
under-development progress towards the next release. The current release is
35+
version [2.24.0](https://github.com/triton-inference-server/server/tree/r22.07)
36+
and corresponds to the 22.07 container release on
37+
[NVIDIA GPU Cloud (NGC)](https://catalog.ngc.nvidia.com/orgs/nvidia/containers/tritonserver).**
38+
39+
----
40+
Triton Inference Server is an open source inference serving software that
41+
streamlines AI inferencing. Triton enables teams to deploy any AI model from
42+
multiple deep learning and machine learning frameworks, including TensorRT,
43+
TensorFlow, PyTorch, ONNX, OpenVINO, Python, RAPIDS FIL, and more. Triton
44+
supports inference across cloud, data center,edge and embedded devices on NVIDIA
45+
GPUs, x86 and ARM CPU, or AWS Inferentia. Triton delivers optimized performance
46+
for many query types, including real time, batched, ensembles and audio/video
47+
streaming.
48+
49+
Major features include:
50+
51+
- [Supports multiple deep learning
52+
frameworks](https://github.com/triton-inference-server/backend#where-can-i-find-all-the-backends-that-are-available-for-triton)
53+
- [Supports multiple machine learning
54+
frameworks](https://github.com/triton-inference-server/fil_backend)
55+
- [Concurrent model
56+
execution](docs/architecture.md#concurrent-model-execution)
57+
- [Dynamic batching](docs/model_configuration.md#dynamic-batcher)
58+
- [Sequence batching](docs/model_configuration.md#sequence-batcher) and
59+
[implicit state management](docs/architecture.md#implicit-state-management)
60+
for stateful models
61+
- Provides [Backend API](https://github.com/triton-inference-server/backend) that
62+
allows adding custom backends and pre/post processing operations
63+
- Model pipelines using
64+
[Ensembling](docs/architecture.md#ensemble-models) or [Business
65+
Logic Scripting
66+
(BLS)](https://github.com/triton-inference-server/python_backend#business-logic-scripting)
67+
- [HTTP/REST and GRPC inference
68+
protocols](docs/inference_protocols.md) based on the community
69+
developed [KServe
70+
protocol](https://github.com/kserve/kserve/tree/master/docs/predict-api/v2)
71+
- A [C API](docs/inference_protocols.md#in-process-triton-server-api) and
72+
[Java API](docs/inference_protocols.md#java-bindings-for-in-process-triton-server-api)
73+
allow Triton to link directly into your application for edge and other in-process use cases
74+
- [Metrics](docs/metrics.md) indicating GPU utilization, server
75+
throughput, server latency, and more
76+
77+
Need enterprise support? NVIDIA global support is available for Triton
78+
Inference Server with the
79+
[NVIDIA AI Enterprise software suite](https://www.nvidia.com/en-us/data-center/products/ai-enterprise/).
80+
81+
## Serve a Model in 3 Easy Steps
82+
83+
```bash
84+
# Step 1: Create the example model repository
85+
git clone -b r22.07 https://github.com/triton-inference-server/server.git
86+
87+
cd server/docs/examples
88+
89+
./fetch_models.sh
90+
91+
# Step 2: Launch triton from the NGC Triton container
92+
docker run --gpus=1 --rm --net=host -v /full/path/to/docs/examples/model_repository:/models nvcr.io/nvidia/tritonserver:22.07-py3 tritonserver --model-repository=/models
93+
94+
# Step 3: In a separate console, launch the image_client example from the NGC Triton SDK container
95+
docker run -it --rm --net=host nvcr.io/nvidia/tritonserver:22.07-py3-sdk
96+
97+
/workspace/install/bin/image_client -m densenet_onnx -c 3 -s INCEPTION /workspace/images/mug.jpg
98+
99+
# Inference should return the following
100+
Image '/workspace/images/mug.jpg':
101+
15.346230 (504) = COFFEE MUG
102+
13.224326 (968) = CUP
103+
10.422965 (505) = COFFEEPOT
104+
```
105+
Please read the [QuickStart](docs/quickstart.md) guide for additional information
106+
regarding this example. The quickstart guide also contains an example of how to launch Triton on [CPU-only systems](docs/quickstart.md#run-on-cpu-only-system).
107+
108+
## Examples and Tutorials
109+
110+
Check out [NVIDIA LaunchPad](https://www.nvidia.com/en-us/data-center/products/ai-enterprise-suite/trial/)
111+
for free access to a set of hands-on labs with Triton Inference Server hosted on
112+
NVIDIA infrastructure.
113+
114+
Specific end-to-end examples for popular models, such as ResNet, BERT, and DLRM
115+
are located in the
116+
[NVIDIA Deep Learning Examples](https://github.com/NVIDIA/DeepLearningExamples)
117+
page on GitHub. The
118+
[NVIDIA Developer Zone](https://developer.nvidia.com/nvidia-triton-inference-server)
119+
contains additional documentation, presentations, and examples.
120+
121+
## Documentation
122+
123+
### Build and Deploy
124+
125+
The recommended way to build and use Triton Inference Server is with Docker
126+
images.
127+
128+
- [Install Triton Inference Server with Docker containers](docs/build.md#building-triton-with-docker) (*Recommended*)
129+
- [Install Triton Inference Server without Docker containers](docs/build.md#building-triton-without-docker)
130+
- [Build a custom Triton Inference Server Docker container](docs/compose.md)
131+
- [Build Triton Inference Server from source](docs/build.md#building-on-unsupported-platforms)
132+
- [Build Triton Inference Server for Windows 10](docs/build.md#building-for-windows-10)
133+
- Examples for deploying Triton Inference Server with Kubernetes and Helm on [GCP](deploy/gcp/README.md),
134+
[AWS](deploy/aws/README.md), and [NVIDIA FleetCommand](deploy/fleetcommand/README.md)
135+
136+
### Using Triton
137+
138+
#### Preparing Models for Triton Inference Server
139+
140+
The first step in using Triton to serve your models is to place one or
141+
more models into a [model repository](docs/model_repository.md). Depending on
142+
the type of the model and on what Triton capabilities you want to enable for
143+
the model, you may need to create a [model
144+
configuration](docs/model_configuration.md) for the model.
145+
146+
- [Add custom operations to Triton if needed by your model](docs/custom_operations.md)
147+
- Enable model pipelining with [Model Ensemble](docs/architecture.md#ensemble-models)
148+
and [Business Logic Scripting (BLS)](https://github.com/triton-inference-server/python_backend#business-logic-scripting)
149+
- Optimize your models setting [scheduling and batching](docs/architecture.md#models-and-schedulers)
150+
parameters and [model instances](docs/model_configuration.md#instance-groups).
151+
- Use the [Model Analyzer tool](https://github.com/triton-inference-server/model_analyzer)
152+
to help optimize your model configuration with profiling
153+
- Learn how to [explicitly manage what models are available by loading and
154+
unloading models](docs/model_management.md)
155+
156+
#### Configure and Use Triton Inference Server
157+
158+
- Read the [Quick Start Guide](docs/quickstart.md) to run Triton Inference
159+
Server on both GPU and CPU
160+
- Triton supports multiple execution engines, called
161+
[backends](https://github.com/triton-inference-server/backend#where-can-i-find-all-the-backends-that-are-available-for-triton), including
162+
[TensorRT](https://github.com/triton-inference-server/tensorrt_backend),
163+
[TensorFlow](https://github.com/triton-inference-server/tensorflow_backend),
164+
[PyTorch](https://github.com/triton-inference-server/pytorch_backend),
165+
[ONNX](https://github.com/triton-inference-server/onnxruntime_backend),
166+
[OpenVINO](https://github.com/triton-inference-server/openvino_backend),
167+
[Python](https://github.com/triton-inference-server/python_backend), and more
168+
- Learn how to [optimize performance](docs/optimization.md) using the
169+
[Performance Analyzer](docs/perf_analyzer.md) and
170+
[Model Analyzer](https://github.com/triton-inference-server/model_analyzer)
171+
- Learn how to [manage loading and unloading models](docs/model_management.md) in
172+
Triton
173+
- Send requests directly to Triton with the [HTTP/REST JSON-based
174+
or gRPC protocols](docs/inference_protocols.md#httprest-and-grpc-protocols)
175+
176+
#### Client Support and Examples
177+
178+
A Triton *client* application sends inference and other requests to Triton. The
179+
[Python and C++ client libraries](https://github.com/triton-inference-server/client)
180+
provide APIs to simplify this communication.
181+
182+
- Review client examples for [C++](https://github.com/triton-inference-server/client/blob/main/src/c%2B%2B/examples),
183+
[Python](https://github.com/triton-inference-server/client/blob/main/src/python/examples),
184+
and [Java](https://github.com/triton-inference-server/client/blob/main/src/java/src/main/java/triton/client/examples)
185+
- Configure [HTTP](https://github.com/triton-inference-server/client#http-options)
186+
and [gRPC](https://github.com/triton-inference-server/client#grpc-options)
187+
client options
188+
- Send input data (e.g. a jpeg image) directly to Triton in the [body of an HTTP
189+
request without any additional metadata](https://github.com/triton-inference-server/server/blob/main/docs/protocol/extension_binary_data.md#raw-binary-request)
190+
191+
### Extend Triton
192+
193+
[Triton Inference Server's architecture](docs/architecture.md) is specifically
194+
designed for modularity and flexibility
195+
196+
- [Customize Triton Inference Server container](docs/compose.md) for your use case
197+
- [Create custom backends](https://github.com/triton-inference-server/backend)
198+
in either [C/C++](https://github.com/triton-inference-server/backend/blob/main/README.md#triton-backend-api)
199+
or [Python](https://github.com/triton-inference-server/python_backend)
200+
- Create [decouple backends and models](docs/decoupled_models.md) that can send
201+
multiple responses for a request or not send any responses for a request
202+
- Use a [Triton repository agent](docs/repository_agents.md) to add functionality
203+
that operates when a model is loaded and unloaded, such as authentication,
204+
decryption, or conversion
205+
- Deploy Triton on [Jetson and JetPack](docs/jetson.md)
206+
- [Use Triton on AWS
207+
Inferentia](https://github.com/triton-inference-server/python_backend/tree/main/inferentia)
208+
209+
### Additional Documentation
210+
211+
- [FAQ](docs/faq.md)
212+
- [User Guide](docs#user-guide)
213+
- [Developer Guide](docs#developer-guide)
214+
- [Release Notes](https://docs.nvidia.com/deeplearning/triton-inference-server/release-notes/index.html)
215+
- [GPU, Driver, and CUDA Support
216+
Matrix](https://docs.nvidia.com/deeplearning/dgx/support-matrix/index.html)
217+
218+
## Contributing
219+
220+
Contributions to Triton Inference Server are more than welcome. To
221+
contribute please review the [contribution
222+
guidelines](CONTRIBUTING.md). If you have a backend, client,
223+
example or similar contribution that is not modifying the core of
224+
Triton, then you should file a PR in the [contrib
225+
repo](https://github.com/triton-inference-server/contrib).
226+
227+
## Reporting problems, asking questions
228+
229+
We appreciate any feedback, questions or bug reporting regarding this project.
230+
When posting [issues in GitHub](https://github.com/triton-inference-server/server/issues),
231+
follow the process outlined in the [Stack Overflow document](https://stackoverflow.com/help/mcve).
232+
Ensure posted examples are:
233+
- minimal – use as little code as possible that still produces the
234+
same problem
235+
- complete – provide all parts needed to reproduce the problem. Check
236+
if you can strip external dependencies and still show the problem. The
237+
less time we spend on reproducing problems the more time we have to
238+
fix it
239+
- verifiable – test the code you're about to provide to make sure it
240+
reproduces the problem. Remove all other problems that are not
241+
related to your request/question.
242+
243+
## For more information
244+
245+
Please refer to the [NVIDIA Developer Triton page](https://developer.nvidia.com/nvidia-triton-inference-server)
246+
for more information.

‎RELEASE.md

+122
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,122 @@
1+
<!--
2+
# Copyright 2018-2022, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
3+
#
4+
# Redistribution and use in source and binary forms, with or without
5+
# modification, are permitted provided that the following conditions
6+
# are met:
7+
# * Redistributions of source code must retain the above copyright
8+
# notice, this list of conditions and the following disclaimer.
9+
# * Redistributions in binary form must reproduce the above copyright
10+
# notice, this list of conditions and the following disclaimer in the
11+
# documentation and/or other materials provided with the distribution.
12+
# * Neither the name of NVIDIA CORPORATION nor the names of its
13+
# contributors may be used to endorse or promote products derived
14+
# from this software without specific prior written permission.
15+
#
16+
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
17+
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18+
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
19+
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
20+
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
21+
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
22+
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
23+
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
24+
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
25+
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
26+
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
27+
-->
28+
29+
# Release Notes for 2.24.0
30+
31+
## New Freatures and Improvements
32+
33+
* [Auto-Complete](https://github.com/triton-inference-server/server/blob/main/docs/model_configuration.md#auto-generated-model-configuration)
34+
is enabled by default. The `--strict-model-config` option has been soft
35+
deprecated, use the new `--disable-auto-complete-config` CLI option instead.
36+
37+
* New example backend demonstrating
38+
[Business Logic Scripting in C++](https://github.com/triton-inference-server/backend/blob/r22.07/examples/backends/bls/README.md).
39+
40+
* Users can provide values for
41+
["init_ops"](https://github.com/triton-inference-server/tensorflow_backend/tree/r22.07#parameters)
42+
in Tensorflow TF1.x GraphDef models through json file.
43+
44+
* New
45+
[asyncio compatible API](https://github.com/triton-inference-server/client#python-asyncio-support-beta)
46+
to the Python GRPC/HTTP APIs.
47+
48+
* Added thread pool to reduce service downtime for concurrently loading models.
49+
The thread pool size is configurable with the new `--model-load-thread-count`
50+
tritonserver option. You can find more information
51+
[here](https://github.com/triton-inference-server/server/blob/main/docs/model_management.md#concurrently-loading-models).
52+
53+
* Model Analyzer now doesn't require `config.pbtxt` file for models that can be
54+
auto-completed in Triton.
55+
56+
* Refer to the 22.07 column of the
57+
[Frameworks Support Matrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html)
58+
for container image versions on which the 22.07 inference server container is
59+
based.
60+
61+
## Known Issues
62+
63+
* JetPack release will be published later in the month in order to align with
64+
JetPack SDK public availability.
65+
66+
* Auto-complete could cause an increase in server start time. To avoid a start
67+
time increase, users should provide the full model configuration.
68+
69+
* When auto-completing some model configs, backends may generate a model config
70+
even though there is not enough metadata (ex. Graphdef models for Tensorflow
71+
Backend). The user will see the model successfully load but fail to inference.
72+
In this case the user should provide the full model configuration for these
73+
models or use the `--disable-auto-complete-config` CLI option to show which
74+
models fail to load.
75+
76+
* Can't do autocomplete for PyTorch models, not enough metadata. Can only verify
77+
that the number of inputs is correct and the input names match what is
78+
specified in the model configuration. No info about number of outputs and
79+
datatypes. Related pytorch bug:
80+
https://github.com/pytorch/pytorch/issues/38273.
81+
82+
* Running inference on multiple TensorRT model instances in Triton may fail
83+
with signal(6). The issue is expected to be fixed in a future release. Details
84+
can be found at https://github.com/triton-inference-server/server/issues/4566.
85+
86+
* Perf Analyzer stability criteria has been changed which may result in
87+
reporting instability for scenarios that were previously considered stable.
88+
This change has been made to improve the accuracy of Perf Analyzer results.
89+
If you observe this message, it can be resolved by increasing the
90+
`--measurement-interval` in the time windows mode or
91+
`--measurement-request-count` in the count windows mode.
92+
93+
* Unlike previously noted, 22.07 is the last release that defaults to
94+
[TensorFlow version 1](https://github.com/triton-inference-server/tensorflow_backend/tree/r22.06#--backend-configtensorflowversionint).
95+
From 22.08 onwards Triton will change the default TensorFlow
96+
version to 2.X.
97+
98+
* Triton PIP wheels for ARM SBSA are not available from PyPI and pip will
99+
install an incorrect Jetson version of Triton for Arm SBSA.
100+
101+
The correct wheel file can be pulled directly from the Arm SBSA SDK image and
102+
manually installed.
103+
104+
* Traced models in PyTorch seem to create overflows when int8 tensor values are
105+
transformed to int32 on the GPU.
106+
107+
Refer to issue https://github.com/pytorch/pytorch/issues/66930 for more
108+
information.
109+
110+
* Triton cannot retrieve GPU metrics with MIG-enabled GPU devices (A100 and A30).
111+
112+
* Triton metrics might not work if the host machine is running a separate DCGM
113+
agent on bare-metal or in a container.
114+
115+
* Starting from 22.02, the Triton container, which uses the 22.02 or above
116+
PyTorch container, will report an error during model loading in the PyTorch
117+
backend when using scripted models that were exported in the legacy format
118+
(using our 19.09 or previous PyTorch NGC containers corresponding to
119+
PyTorch 1.2.0 or previous releases).
120+
121+
To load the model successfully in Triton, you need to export the model again
122+
by using a recent version of PyTorch.

0 commit comments

Comments
 (0)