<!-- # Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # * Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # * Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # * Neither the name of NVIDIA CORPORATION nor the names of its # contributors may be used to endorse or promote products derived # from this software without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY # EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR # PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR # CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, # PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR # PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY # OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. --> <!-- https://linear.app/issue/TRI-417 --> # Triton Inference Server The Triton Inference Server provides a cloud inferencing solution optimized for both CPUs and GPUs. The server provides an inference service via an HTTP or GRPC endpoint, allowing remote clients to request inferencing for any model being managed by the server. For edge deployments, Triton Server is also available as a shared library with an API that allows the full functionality of the server to be included directly in an application. <details> <summary><h2>New Features and Improvements</h2></summary> * Fixed an issue with Triton Server’s Sagemaker Service which could result in a server crash resulting from a race condition, caused by unprotected access to the list of models. * Extended the set of accelerated PyTorch libraries included with the Triton PyTorch backend. * Upgraded Triton Client’s Golang dependencies to latest stable versions to resolve known issues with the previous version of the dependencies. * The OpenAI-compatible frontend has transitioned from beta to a stable release. * Added `echo` request parameter for TensorRT-LLM and Python backends to OpenAI-compatible API frontend `v1/completions` endpoint. * Enabled OpenAI-compatible API frontend multi-LoRA support for TensorRT-LLM backend. * Backends can now implement the new `TRITONBACKEND_ModelInstanceReady` function to report accurate model readiness status. * Updated the Python backend to accurately report model readiness. </details> <details> <summary><h2>Known Issues</h2></summary> * The error `'__init__(): incompatible function arguments` may occur when using TensorRT-LLM backend python models. To suppress the error temporarily, set input tensor `stream` with a boolean value explicitly in the request header. * Since 25.10, vLLM backend uses V1 engine by default. You might see invalid characters in logprobs output and the bug has been reported to the vLLM team. * PyTorch backend supports PyTorch 2.0 with the limitation that models must be provided as a serialized model file (aka [‘model.pt’](http://model.pt)). Please see Triton PyTorch Backend documentation for details. * vLLM's v0 API and Ray are affected by vulnerabilities. Users should consider their own architecture and mitigation steps which may include but should not be limited to: * Do not expose Ray executors and vLLM hosts to a network where any untrusted connections might reach the host. * Ensure that only the other vLLM hosts are able to connect to the TCP port used for the XPUB socket. Note that the port used is random. * Perf Analyzer is no longer part of the “client” released archive and can be installed separately using `pip install perf-analyzer`. * When using Valgrind or other leak detection tools on AGX-Thor or DGX-Spark systems, you might see memory leaks attributed to NvRmGpuLibOpen. The root cause has been identified and fixed in CUDA. * Valgrind or other memory leak detection tools may occasionally report leaks related to DCGM. These reports are intermittent and often disappear on retry. The root cause is under investigation. * CuPy has issues with the CUDA 13 Device API in multithreaded contexts. Avoid using tritonclient cuda_shared_memory APIs in multithreaded environments until fixed by CuPy. * TensorRT calibration cache may require size adjustment in some cases, which was observed for the IGX platform. * The core Python binding may incur an additional D2H and H2D copy if the backend and frontend both specify device memory to be used for response tensors. * A segmentation fault related to DCGM and NSCQ may be encountered during server shutdown on NVSwitch systems. A possible workaround for this issue is to disable the collection of GPU metrics `tritonserver --allow-gpu-metrics false ...` * When using TensorRT models, if auto-complete configuration is disabled and `is_non_linear_format_io:true` for [reformat-free tensors](https://github.com/triton-inference-server/server/blob/r24.08/docs/user_guide/model_configuration.md#non-linear-io-formats) is not provided in the model configuration, the model may not load successfully. * When using Python models in [decoupled mode](https://github.com/triton-inference-server/python_backend/tree/main?tab=readme-ov-file#decoupled-mode), users need to ensure that the `ResponseSender` goes out of scope or is properly cleaned up before unloading the model to guarantee that the unloading process executes correctly. * Triton Inference Server with vLLM backend currently does not support running vLLM models with tensor parallelism sizes greater than 1 and the default "distributed_executor_backend" setting when using explicit model control mode. In attempt to load a vllm model (tp > 1) in explicit mode, users could potentially see failure at `initialize` step: `could not acquire lock for <_io.BufferedWriter name='<stdout>'> at interpreter shutdown, possibly due to daemon threads`. For the default model control mode, after server shutdown, vllm related sub-processes are not killed. Related vllm issue: https://github.com/vllm-project/vllm/issues/6766 . Please specify "distributed_executor_backend":"ray" in the `model.json` when deploying vllm models with tensor parallelism > 1. * When loading models with file override, multiple model configuration files are not supported. Users must provide the model configuration by setting parameter `"config" : "<JSON>"` instead of custom configuration file in the following format: `"file:configs/<model-config-name>.pbtxt" : "<base64-encoded-file-content>"`. * TensorRT-LLM [backend](https://github.com/triton-inference-server/tensorrtllm_backend) provides limited support of Triton extensions and features. * The TensorRT-LLM backend may core dump on server shutdown. This impacts server teardown only and will not impact inferencing. * The Java CAPI is known to have intermittent segfaults. * Some systems which implement `malloc()` may not release memory back to the operating system right away causing a false memory leak. This can be mitigated by using a different malloc implementation. `TCMalloc` and `jemalloc` are installed in the Triton container and can be [used by specifying the library in LD_PRELOAD](https://github.com/triton-inference-server/server/blob/r25.01/docs/user_guide/model_management.md). NVIDIA recommends experimenting with both `tcmalloc` and `jemalloc` to determine which one works better for your use case. * Auto-complete may cause an increase in server start time. To avoid a start time increase, users can provide the full model configuration and launch the server with `--disable-auto-complete-config`. * Auto-complete does not support PyTorch models due to lack of metadata in the model. It can only verify that the number of inputs and the input names matches what is specified in the model configuration. There is no model metadata about the number of outputs and datatypes. Related PyTorch bug:[ https://github.com/pytorch/pytorch/issues/38273](https://github.com/pytorch/pytorch/issues/38273) * Triton Client PIP wheels for ARM SBSA are not available from PyPI and pip will install an incorrect Jetson version of Triton Client library for Arm SBSA. The correct client wheel file can be pulled directly from the Arm SBSA SDK image and manually installed. * Traced models in PyTorch seem to create overflows when int8 tensor values are transformed to int32 on the GPU. Refer to [pytorch/pytorch#66930](https://github.com/pytorch/pytorch/issues/66930) for more information. * Triton cannot retrieve GPU metrics with [MIG-enabled GPU devices](https://docs.nvidia.com/datacenter/tesla/mig-user-guide/index.html#supported-gpus). * Triton metrics might not work if the host machine is running a separate DCGM agent on bare-metal or in a container. * When cloud storage (AWS, GCS, AZURE) is used as a model repository and a model has multiple versions, Triton creates an extra local copy of the cloud model’s folder in the temporary directory, which is deleted upon server’s shutdown. * Python backend support for Windows is limited and does not currently support the following features: - GPU tensors - CPU and GPU-related metrics - Custom execution environments - The model load/unload APIs </details> <details> <summary><h2>Client Libraries and Examples</h2></summary> Ubuntu 24.04 builds of the client libraries and examples are included in this release in the attached `v2.64.0_ubuntu2404.clients.tar.gz` file. The SDK is also available for as an Ubuntu 24.04 based [NGC Container](https://ngc.nvidia.com/catalog/containers/nvidia:tritonserver/tags). The SDK container includes the client libraries and examples, Performance Analyzer and Model Analyzer. Some components are also available in the tritonclient pip package. See [Getting the Client Libraries](https://github.com/triton-inference-server/client/tree/r25.12#getting-the-client-libraries-and-examples) for more information on each of these options. </details> <details> <summary><h2>Jetson iGPU Support</h2></summary> A release of Triton for [IGX](https://www.nvidia.com/en-us/edge-computing/products/igx/) is provided in the attached tar file: [`tritonserver2.64.0-igpu.tar`](https://github.com/triton-inference-server/server/releases/download/v2.64.0/tritonserver2.64.0-igpu.tar). * This release supports **CUDA** `12.9`, **TensorRT** `10.11.0.33`, **Onnx Runtime** `1.23.2`, **PyTorch** [`2.10.0a0+b4e4ee8`](https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform-release-notes/pytorch-jetson-rel.html), **Python** `3.12` and as well as _ensembles_. * ONNX Runtime backend does not support the OpenVINO and TensorRT execution providers. The CUDA execution provider is in Beta. * System shared memory is supported on Jetson. CUDA shared memory is not supported. * GPU metrics, GCS storage, S3 storage and Azure storage are not supported. The tar file contains the Triton server executable and shared libraries and also the C++ and Python client libraries and examples. For more information on how to install and use Triton on JetPack refer to [`jetson.md`](https://github.com/triton-inference-server/server/blob/r25.10/docs/user_guide/jetson.md). The wheel for the Python client library is present in the tar file and can be installed by running the following command: ``` python3 -m pip install --upgrade clients/python/tritonclient-2.64.0-py3-none-manylinux2014_aarch64.whl[all] ``` </details> <details> <summary><h2>Jetson AGX Systems Support</h2></summary> A release of Triton for [AGX Systems](https://www.nvidia.com/en-us/deep-learning-ai/products/agx-systems) is provided in the attached tar file: [`tritonserver2.64.0-agx.tar`](https://github.com/triton-inference-server/server/releases/download/v2.64.0/tritonserver2.64.0-agx.tar). * This release supports **CUDA** `13.1`, **TensorRT** `10.14.1.48`, **Onnx Runtime** `1.23.2`, **PyTorch** [`2.10.0a0+b4e4ee8`](https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform-release-notes/pytorch-jetson-rel.html), **Python** `3.12` and as well as _ensembles_. * This package is a subset of `nvcr.io/nvidia/tritonserver:25.12-py3` ARM container image assets it. </details> <details> <summary><h2>Triton TRT-LLM Container Support Matrix</h2></summary> The Triton TensorRT-LLM container is built from the 25.10 image [`nvcr.io/nvidia/tritonserver:25.10-py3-min`](http://nvcr.io/nvidia/tritonserver:25.10-py3-min). Please refer to the [support matrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html) and [compatibility.md](https://github.com/triton-inference-server/server/blob/v2.64.0/docs/introduction/compatibility.md#container-name-trtllm-python-py3) for all dependency versions related to 25.10. However, the packages listed below have different versions than those specified in the support matrix. | Dependency | Version | | :------------: | :---------------: | | TensorRT-LLM | 1.1.0 | | TensorRT | 10.13.3.9 | </details> <details> <summary><h2>Windows Support</h2></summary> > [!NOTE] > There is no Windows release for 25.12, the latest release is [25.01](https://github.com/triton-inference-server/server/releases/tag/v2.54.0). </details> <details> <summary><h2>ManyLinux Assets (early access)</h2></summary> This release was compiled with AlmaLinux 8.9 based out of `manylinux_2_28` and can be used on RHEL8 and later versions. See the included README.md for complete details about installation, verification, and support. This release supports CUDA 13, TensorRT 10.14.1.48, Onnx Runtime 1.23.2, PyTorch 2.10.0a0+b4e4ee8, Python 3.12 and supports ensembles. Some optional backend features such as the PyTorch backend's TorchTRT extension are not currently supported. </details>

<!-- TRI-154 --> # Triton Inference Server The Triton Inference Server provides a cloud inferencing solution optimized for both CPUs and GPUs. The server provides an inference service via an HTTP or GRPC endpoint, allowing remote clients to request inferencing for any model being managed by the server. For edge deployments, Triton Server is also available as a shared library with an API that allows the full functionality of the server to be included directly in an application. <details> <summary><h2>New Features and Improvements</h2></summary> * Fixed a server crash issue caused by specially crafted request messages sent to `/v2/models/<model_name>/infer`. * Fixed a server crash issue caused by incorrect handling of malformed HTTP requests. </details> <details> <summary><h2>Known Issues</h2></summary> * Triton python package uses outdated dependency `starlette` package version. * Since 25.10, vLLM backend uses V1 engine by default. You might see invalid characters in logprobs output and the bug has been reported to the vLLM team. * Enabling vLLM metrics during inferences causes the engine to crash. * PyTorch backend supports PyTorch 2.0 with the limitation that models must be provided as a serialized model file (aka ‘model.pt’). Please see Triton PyTorch Backend documentation for details. * vLLM's v0 API and Ray are affected by vulnerabilities. Users should consider their own architecture and mitigation steps which may include but should not be limited to: * Do not expose Ray executors and vLLM hosts to a network where any untrusted connections might reach the host. * Ensure that only the other vLLM hosts are able to connect to the TCP port used for the XPUB socket. Note that the port used is random. * Perf Analyzer is no longer part of the “client” released archive and can be installed separately using `pip install perf-analyzer`. * When using Valgrind or other leak detection tools on AGX-Thor or DGX-Spark systems, you might see memory leaks attributed to NvRmGpuLibOpen. The root cause has been identified and fixed in CUDA. * Valgrind or other memory leak detection tools may occasionally report leaks related to DCGM. These reports are intermittent and often disappear on retry. The root cause is under investigation. * CuPy has issues with the CUDA 13 Device API in multithreaded contexts. Avoid using tritonclient cuda_shared_memory APIs in multithreaded environments until fixed by CuPy. * TensorRT calibration cache may require size adjustment in some cases, which was observed for the IGX platform. * The core Python binding may incur an additional D2H and H2D copy if the backend and frontend both specify device memory to be used for response tensors. * A segmentation fault related to DCGM and NSCQ may be encountered during server shutdown on NVSwitch systems. A possible workaround for this issue is to disable the collection of GPU metrics `tritonserver --allow-gpu-metrics false ...` * When using TensorRT models, if auto-complete configuration is disabled and `is_non_linear_format_io:true` for [reformat-free tensors](https://github.com/triton-inference-server/server/blob/r24.08/docs/user_guide/model_configuration.md#non-linear-io-formats) is not provided in the model configuration, the model may not load successfully. * When using Python models in [decoupled mode](https://github.com/triton-inference-server/python_backend/tree/main?tab=readme-ov-file#decoupled-mode), users need to ensure that the `ResponseSender` goes out of scope or is properly cleaned up before unloading the model to guarantee that the unloading process executes correctly. * Triton Inference Server with vLLM backend currently does not support running vLLM models with tensor parallelism sizes greater than 1 and the default "distributed_executor_backend" setting when using explicit model control mode. In attempt to load a vllm model (tp > 1) in explicit mode, users could potentially see failure at `initialize` step: `could not acquire lock for <_io.BufferedWriter name='<stdout>'> at interpreter shutdown, possibly due to daemon threads`. For the default model control mode, after server shutdown, vllm related sub-processes are not killed. Related vllm issue: https://github.com/vllm-project/vllm/issues/6766 . Please specify "distributed_executor_backend":"ray" in the `model.json` when deploying vllm models with tensor parallelism > 1. * When loading models with file override, multiple model configuration files are not supported. Users must provide the model configuration by setting parameter `"config" : "<JSON>"` instead of custom configuration file in the following format: `"file:configs/<model-config-name>.pbtxt" : "<base64-encoded-file-content>"`. * TensorRT-LLM [backend](https://github.com/triton-inference-server/tensorrtllm_backend) provides limited support of Triton extensions and features. * The TensorRT-LLM backend may core dump on server shutdown. This impacts server teardown only and will not impact inferencing. * The Java CAPI is known to have intermittent segfaults. * Some systems which implement `malloc()` may not release memory back to the operating system right away causing a false memory leak. This can be mitigated by using a different malloc implementation. `TCMalloc` and `jemalloc` are installed in the Triton container and can be [used by specifying the library in LD_PRELOAD](https://github.com/triton-inference-server/server/blob/r25.01/docs/user_guide/model_management.md). NVIDIA recommends experimenting with both `tcmalloc` and `jemalloc` to determine which one works better for your use case. * Auto-complete may cause an increase in server start time. To avoid a start time increase, users can provide the full model configuration and launch the server with `--disable-auto-complete-config`. * Auto-complete does not support PyTorch models due to lack of metadata in the model. It can only verify that the number of inputs and the input names matches what is specified in the model configuration. There is no model metadata about the number of outputs and datatypes. Related PyTorch bug:[ https://github.com/pytorch/pytorch/issues/38273](https://github.com/pytorch/pytorch/issues/38273) * Triton Client PIP wheels for ARM SBSA are not available from PyPI and pip will install an incorrect Jetson version of Triton Client library for Arm SBSA. The correct client wheel file can be pulled directly from the Arm SBSA SDK image and manually installed. * Traced models in PyTorch seem to create overflows when int8 tensor values are transformed to int32 on the GPU. Refer to [pytorch/pytorch#66930](https://github.com/pytorch/pytorch/issues/66930) for more information. * Triton cannot retrieve GPU metrics with [MIG-enabled GPU devices](https://docs.nvidia.com/datacenter/tesla/mig-user-guide/index.html#supported-gpus). * Triton metrics might not work if the host machine is running a separate DCGM agent on bare-metal or in a container. * When cloud storage (AWS, GCS, AZURE) is used as a model repository and a model has multiple versions, Triton creates an extra local copy of the cloud model’s folder in the temporary directory, which is deleted upon server’s shutdown. * Python backend support for Windows is limited and does not currently support the following features: - GPU tensors - CPU and GPU-related metrics - Custom execution environments - The model load/unload APIs </details> <details> <summary><h2>Client Libraries and Examples</h2></summary> Ubuntu 24.04 builds of the client libraries and examples are included in this release in the attached `v2.62.0_ubuntu2404.clients.tar.gz` file. The SDK is also available for as an Ubuntu 24.04 based [NGC Container](https://ngc.nvidia.com/catalog/containers/nvidia:tritonserver/tags). The SDK container includes the client libraries and examples, Performance Analyzer and Model Analyzer. Some components are also available in the tritonclient pip package. See [Getting the Client Libraries](https://github.com/triton-inference-server/client/tree/r25.10#getting-the-client-libraries-and-examples) for more information on each of these options. </details> <details> <summary><h2>Windows Support</h2></summary> > [!NOTE] > There is no Windows release for 25.08, the latest release is [25.01](https://github.com/triton-inference-server/server/releases/tag/v2.54.0). </details> <details> <summary><h2>Jetson iGPU Support</h2></summary> A release of Triton for [IGX](https://www.nvidia.com/en-us/edge-computing/products/igx/) is provided in the attached tar file: [`tritonserver2.62.0-igpu.tar`](https://github.com/triton-inference-server/server/releases/download/v2.62.0/tritonserver2.62.0-igpu.tar). * This release supports **CUDA** `12.9`, **TensorRT** `10.11.0.33`, **Onnx Runtime** `1.23.1`, **PyTorch** [`2.9.0a0+145a3a7`](https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform-release-notes/pytorch-jetson-rel.html), **Python** `3.12` and as well as _ensembles_. * ONNX Runtime backend does not support the OpenVINO and TensorRT execution providers. The CUDA execution provider is in Beta. * System shared memory is supported on Jetson. CUDA shared memory is not supported. * GPU metrics, GCS storage, S3 storage and Azure storage are not supported. The tar file contains the Triton server executable and shared libraries and also the C++ and Python client libraries and examples. For more information on how to install and use Triton on JetPack refer to [`jetson.md`](https://github.com/triton-inference-server/server/blob/r25.10/docs/user_guide/jetson.md). The wheel for the Python client library is present in the tar file and can be installed by running the following command: ``` python3 -m pip install --upgrade clients/python/tritonclient-2.62.0-py3-none-manylinux2014_aarch64.whl[all] ``` </details> <details> <summary><h2>Jetson AGX Systems Support</h2></summary> A release of Triton for [AGX Systems](https://www.nvidia.com/en-us/deep-learning-ai/products/agx-systems) is provided in the attached tar file: [`tritonserver2.62.0-agx.tar`](https://github.com/triton-inference-server/server/releases/download/v2.62.0/tritonserver2.62.0-agx.tar). * This release supports **CUDA** `13.0`, **TensorRT** `10.13.3.9`, **Onnx Runtime** `1.23.1`, **PyTorch** [`2.9.0a0+50eac81`](https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform-release-notes/pytorch-jetson-rel.html), **Python** `3.12` and as well as _ensembles_. * This package is a subset of `nvcr.io/nvidia/tritonserver:25.10-py3` ARM container image assets it. </details> <details> <summary><h2>Triton TRT-LLM Container Support Matrix</h2></summary> The Triton TensorRT-LLM container is built from the 25.06 image [`nvcr.io/nvidia/tritonserver:25.06-py3-min`](http://nvcr.io/nvidia/tritonserver:25.06-py3-min). Please refer to the [support matrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html) and [compatibility.md](https://github.com/triton-inference-server/server/blob/v2.62.0/docs/introduction/compatibility.md#container-name-trtllm-python-py3) for all dependency versions related to 25.06. However, the packages listed below have different versions than those specified in the support matrix. | Dependency | Version | | :------------: | :---------------: | | TensorRT-LLM | 1.0.0 | | TensorRT | 10.11.0.33 | </details>

<!-- TPRD-1607 --> # Triton Inference Server The Triton Inference Server provides a cloud inferencing solution optimized for both CPUs and GPUs. The server provides an inference service via an HTTP or GRPC endpoint, allowing remote clients to request inferencing for any model being managed by the server. For edge deployments, Triton Server is also available as a shared library with an API that allows the full functionality of the server to be included directly in an application. <details> <summary><h2>New Features and Improvements</h2></summary> * Added CUDA 13 support. </details> <details> <summary><h2>Known Issues</h2></summary> * Triton ONNX Runtime Backend build uses [microsoft/onnxruntime/commit/1d1712fdaf](https://github.com/microsoft/onnxruntime/commit/1d1712fdafb9e61b2d6d033c4433c1033395d7e7) and may have some limitations on DGX Spark hardware which will be addressed in future versions. * CuPy has issues with the CUDA 13 Device API in multithreaded contexts. Avoid using tritonclient cuda_shared_memory APIs in multithreaded environments until fixed by CuPy * CuPy does not support CUDA 13 at the time of writing. Issues may be encountered when using CuPy before it officially supports CUDA 13, see https://github.com/triton-inference-server/server/tree/r25.08/python/openai#pre-requisites for more details * TensorRT calibration cache may require size adjustment in some cases, which was observed for the IGX platform. * The core Python binding may incur an additional D2H and H2D copy if the backend and frontend both specify device memory to be used for response tensors. * A segmentation fault related to DCGM and NSCQ may be encountered during server shutdown on NVSwitch systems. A possible workaround for this issue is to disable the collection of GPU metrics `tritonserver --allow-gpu-metrics false ...` * When using TensorRT models, if auto-complete configuration is disabled and `is_non_linear_format_io:true` for [reformat-free tensors](https://github.com/triton-inference-server/server/blob/r24.08/docs/user_guide/model_configuration.md#non-linear-io-formats) is not provided in the model configuration, the model may not load successfully. * When using Python models in [decoupled mode](https://github.com/triton-inference-server/python_backend/tree/main?tab=readme-ov-file#decoupled-mode), users need to ensure that the `ResponseSender` goes out of scope or is properly cleaned up before unloading the model to guarantee that the unloading process executes correctly. * Restart support was temporarily removed for Python models. * Triton Inference Server with vLLM backend currently does not support running vLLM models with tensor parallelism sizes greater than 1 and the default "distributed_executor_backend" setting when using explicit model control mode. In attempt to load a vllm model (tp > 1) in explicit mode, users could potentially see failure at `initialize` step: `could not acquire lock for <_io.BufferedWriter name='<stdout>'> at interpreter shutdown, possibly due to daemon threads`. For the default model control mode, after server shutdown, vllm related sub-processes are not killed. Related vllm issue: https://github.com/vllm-project/vllm/issues/6766 . Please specify "distributed_executor_backend":"ray" in the `model.json` when deploying vllm models with tensor parallelism > 1. * When loading models with file override, multiple model configuration files are not supported. Users must provide the model configuration by setting parameter `"config" : "<JSON>"` instead of custom configuration file in the following format: `"file:configs/<model-config-name>.pbtxt" : "<base64-encoded-file-content>"`. * TensorRT-LLM [backend](https://github.com/triton-inference-server/tensorrtllm_backend) provides limited support of Triton extensions and features. * The TensorRT-LLM backend may core dump on server shutdown. This impacts server teardown only and will not impact inferencing. * The Java CAPI is known to have intermittent segfaults. * Some systems which implement `malloc()` may not release memory back to the operating system right away causing a false memory leak. This can be mitigated by using a different malloc implementation. `TCMalloc` and `jemalloc` are installed in the Triton container and can be [used by specifying the library in LD_PRELOAD](https://github.com/triton-inference-server/server/blob/r25.01/docs/user_guide/model_management.md). NVIDIA recommends experimenting with both `tcmalloc` and `jemalloc` to determine which one works better for your use case. * Auto-complete may cause an increase in server start time. To avoid a start time increase, users can provide the full model configuration and launch the server with `--disable-auto-complete-config`. * Auto-complete does not support PyTorch models due to lack of metadata in the model. It can only verify that the number of inputs and the input names matches what is specified in the model configuration. There is no model metadata about the number of outputs and datatypes. Related PyTorch bug:[ https://github.com/pytorch/pytorch/issues/38273](https://github.com/pytorch/pytorch/issues/38273) * Triton Client PIP wheels for ARM SBSA are not available from PyPI and pip will install an incorrect Jetson version of Triton Client library for Arm SBSA. The correct client wheel file can be pulled directly from the Arm SBSA SDK image and manually installed. * Traced models in PyTorch seem to create overflows when int8 tensor values are transformed to int32 on the GPU. Refer to [pytorch/pytorch#66930](https://github.com/pytorch/pytorch/issues/66930) for more information. * Triton cannot retrieve GPU metrics with [MIG-enabled GPU devices](https://docs.nvidia.com/datacenter/tesla/mig-user-guide/index.html#supported-gpus). * Triton metrics might not work if the host machine is running a separate DCGM agent on bare-metal or in a container. * When cloud storage (AWS, GCS, AZURE) is used as a model repository and a model has multiple versions, Triton creates an extra local copy of the cloud model’s folder in the temporary directory, which is deleted upon server’s shutdown. * Python backend support for Windows is limited and does not currently support the following features: - GPU tensors - CPU and GPU-related metrics - Custom execution environments - The model load/unload APIs </details> <details> <summary><h2>Client Libraries and Examples</h2></summary> Ubuntu 24.04 builds of the client libraries and examples are included in this release in the attached `v2.60.0_ubuntu2404.clients.tar.gz` file. The SDK is also available for as an Ubuntu 24.04 based [NGC Container](https://ngc.nvidia.com/catalog/containers/nvidia:tritonserver/tags). The SDK container includes the client libraries and examples, Performance Analyzer and Model Analyzer. Some components are also available in the tritonclient pip package. See [Getting the Client Libraries](https://github.com/triton-inference-server/client/tree/r25.08#getting-the-client-libraries-and-examples) for more information on each of these options. </details> <details> <summary><h2>Windows Support</h2></summary> > [!NOTE] > There is no Windows release for 25.08, the latest release is [25.01](https://github.com/triton-inference-server/server/releases/tag/v2.54.0). </details> <details> <summary><h2>Jetson iGPU Support</h2></summary> A release of Triton for [IGX](https://www.nvidia.com/en-us/edge-computing/products/igx/) is provided in the attached tar file: [`tritonserver2.60.0-igpu.tar`](https://github.com/triton-inference-server/server/releases/download/v2.60.0/tritonserver2.60.0-igpu.tar). * This release supports **CUDA** `12.9`, **TensorRT** `10.11.0.33`, **Onnx Runtime** `1.23.0+1d1712fdaf`, **PyTorch** [`2.8.0a+34c6371d24`](https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform-release-notes/pytorch-jetson-rel.html), **Python** `3.12` and as well as _ensembles_. * ONNX Runtime backend does not support the OpenVINO and TensorRT execution providers. The CUDA execution provider is in Beta. * System shared memory is supported on Jetson. CUDA shared memory is not supported. * GPU metrics, GCS storage, S3 storage and Azure storage are not supported. The tar file contains the Triton server executable and shared libraries and also the C++ and Python client libraries and examples. For more information on how to install and use Triton on JetPack refer to [`jetson.md`](https://github.com/triton-inference-server/server/blob/r25.07/docs/user_guide/jetson.md). The wheel for the Python client library is present in the tar file and can be installed by running the following command: ``` python3 -m pip install --upgrade clients/python/tritonclient-2.59.0-py3-none-manylinux2014_aarch64.whl[all] ``` </details> <details> <summary><h2>Jetson AGX Systems Support</h2></summary> A release of Triton for [AGX Systems](https://www.nvidia.com/en-us/deep-learning-ai/products/agx-systems) is provided in the attached tar file: [`tritonserver2.60.0-agx.tar`](https://github.com/triton-inference-server/server/releases/download/v2.60.0/tritonserver2.60.0-agx.tar). > [!NOTE] > There is no Jetson AGX release for 25.08, requires DCGM version 4 to be installed in order to use GPU metrics. > Please use following command to install DCGM 4: > ``` > curl -o /tmp/cuda-keyring.deb \ > https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/x86_64/cuda-keyring_1.1-1_all.deb \ > && apt install /tmp/cuda-keyring.deb \ > && rm /tmp/cuda-keyring.deb \ > && apt update \ > && apt install --yes --no-install-recommends \ > datacenter-gpu-manager-4-core=1:4.4.0-1 > ``` * This release supports **CUDA** `13.0`, **TensorRT** `10.13.2.6`, **Onnx Runtime** `1.23.0+1d1712fdaf`, **PyTorch** [`2.8.0a0+34c6371`](https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform-release-notes/pytorch-jetson-rel.html), **Python** `3.12` and as well as _ensembles_. * This package is a subset of `nvcr.io/nvidia/tritonserver:25.08-py3` ARM container image assets it. </details> <details> <summary><h2>Triton TRT-LLM Container Support Matrix</h2></summary> The Triton TensorRT-LLM container is built from the 25.04 image [`nvcr.io/nvidia/tritonserver:25.06-py3-min`](http://nvcr.io/nvidia/tritonserver:25.06-py3-min). Please refer to the [support matrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html) and [compatibility.md](https://github.com/triton-inference-server/server/blob/v2.60.0/docs/introduction/compatibility.md#container-name-trtllm-python-py3) for all dependency versions related to 25.04. However, the packages listed below have different versions than those specified in the support matrix. | Dependency | Version | | :------------: | :---------------: | | TensorRT-LLM | 0.21.0 | | TensorRT | 10.11.0.33 | </details>

<!-- TPRD-1564 --> # Triton Inference Server The Triton Inference Server provides a cloud inferencing solution optimized for both CPUs and GPUs. The server provides an inference service via an HTTP or GRPC endpoint, allowing remote clients to request inferencing for any model being managed by the server. For edge deployments, Triton Server is also available as a shared library with an API that allows the full functionality of the server to be included directly in an application. <details> <summary><h2>New Features and Improvements</h2></summary> * Improved ensemble model performance in scenarios that allow out-of-order responses by increasing maximum throughput and reducing latency. </details> <details> <summary><h2>Known Issues</h2></summary> * TensorRT calibration cache may require size adjustment in some cases, which was observed for the IGX platform. * The core Python binding may incur an additional D2H and H2D copy if the backend and frontend both specify device memory to be used for response tensors. * A segmentation fault related to DCGM and NSCQ may be encountered during server shutdown on NVSwitch systems. A possible workaround for this issue is to disable the collection of GPU metrics `tritonserver --allow-gpu-metrics false ...` * vLLM backend currently does not take advantage of the [vLLM v0.6](https://blog.vllm.ai/2024/09/05/perf-update.html) performance improvement when metrics are enabled. * When using TensorRT models, if auto-complete configuration is disabled and `is_non_linear_format_io:true` for [reformat-free tensors](https://github.com/triton-inference-server/server/blob/r24.08/docs/user_guide/model_configuration.md#non-linear-io-formats) is not provided in the model configuration, the model may not load successfully. * When using Python models in [decoupled mode](https://github.com/triton-inference-server/python_backend/tree/main?tab=readme-ov-file#decoupled-mode), users need to ensure that the `ResponseSender` goes out of scope or is properly cleaned up before unloading the model to guarantee that the unloading process executes correctly. * Restart support was temporarily removed for Python models. * Triton Inference Server with vLLM backend currently does not support running vLLM models with tensor parallelism sizes greater than 1 and the default "distributed_executor_backend" setting when using explicit model control mode. In attempt to load a vllm model (tp > 1) in explicit mode, users could potentially see failure at `initialize` step: `could not acquire lock for <_io.BufferedWriter name='<stdout>'> at interpreter shutdown, possibly due to daemon threads`. For the default model control mode, after server shutdown, vllm related sub-processes are not killed. Related vllm issue: https://github.com/vllm-project/vllm/issues/6766 . Please specify "distributed_executor_backend":"ray" in the `model.json` when deploying vllm models with tensor parallelism > 1. * When loading models with file override, multiple model configuration files are not supported. Users must provide the model configuration by setting parameter `"config" : "<JSON>"` instead of custom configuration file in the following format: `"file:configs/<model-config-name>.pbtxt" : "<base64-encoded-file-content>"`. * TensorRT-LLM [backend](https://github.com/triton-inference-server/tensorrtllm_backend) provides limited support of Triton extensions and features. * The TensorRT-LLM backend may core dump on server shutdown. This impacts server teardown only and will not impact inferencing. * The Java CAPI is known to have intermittent segfaults. * Some systems which implement `malloc()` may not release memory back to the operating system right away causing a false memory leak. This can be mitigated by using a different malloc implementation. `TCMalloc` and `jemalloc` are installed in the Triton container and can be [used by specifying the library in LD_PRELOAD](https://github.com/triton-inference-server/server/blob/r25.01/docs/user_guide/model_management.md). NVIDIA recommends experimenting with both `tcmalloc` and `jemalloc` to determine which one works better for your use case. * Auto-complete may cause an increase in server start time. To avoid a start time increase, users can provide the full model configuration and launch the server with `--disable-auto-complete-config`. * Auto-complete does not support PyTorch models due to lack of metadata in the model. It can only verify that the number of inputs and the input names matches what is specified in the model configuration. There is no model metadata about the number of outputs and datatypes. Related PyTorch bug:[ https://github.com/pytorch/pytorch/issues/38273](https://github.com/pytorch/pytorch/issues/38273) * Triton Client PIP wheels for ARM SBSA are not available from PyPI and pip will install an incorrect Jetson version of Triton Client library for Arm SBSA. The correct client wheel file can be pulled directly from the Arm SBSA SDK image and manually installed. * Traced models in PyTorch seem to create overflows when int8 tensor values are transformed to int32 on the GPU. Refer to [pytorch/pytorch#66930](https://github.com/pytorch/pytorch/issues/66930) for more information. * Triton cannot retrieve GPU metrics with [MIG-enabled GPU devices](https://docs.nvidia.com/datacenter/tesla/mig-user-guide/index.html#supported-gpus). * Triton metrics might not work if the host machine is running a separate DCGM agent on bare-metal or in a container. * When cloud storage (AWS, GCS, AZURE) is used as a model repository and a model has multiple versions, Triton creates an extra local copy of the cloud model’s folder in the temporary directory, which is deleted upon server’s shutdown. * Python backend support for Windows is limited and does not currently support the following features: - GPU tensors - CPU and GPU-related metrics - Custom execution environments - The model load/unload APIs </details> <details> <summary><h2>Client Libraries and Examples</h2></summary> Ubuntu 24.04 builds of the client libraries and examples are included in this release in the attached `v2.59.0_ubuntu2404.clients.tar.gz` file. The SDK is also available for as an Ubuntu 24.04 based [NGC Container](https://ngc.nvidia.com/catalog/containers/nvidia:tritonserver/tags). The SDK container includes the client libraries and examples, Performance Analyzer and Model Analyzer. Some components are also available in the tritonclient pip package. See [Getting the Client Libraries](https://github.com/triton-inference-server/client/tree/r25.05#getting-the-client-libraries-and-examples) for more information on each of these options. </details> <details> <summary><h2>Windows Support</h2></summary> > [!NOTE] > There is no Windows release for 25.06, the latest release is [25.02](https://github.com/triton-inference-server/server/releases/tag/v2.55.0). </details> <details> <summary><h2>Jetson iGPU Support</h2></summary> A release of Triton for [IGX](https://www.nvidia.com/en-us/edge-computing/products/igx/) is provided in the attached tar file: [`tritonserver2.59.0-igpu.tar`](https://github.com/triton-inference-server/server/releases/download/v2.59.0/tritonserver2.59.0-igpu.tar). * This release supports **TensorRT** `10.11.0.33`, **Onnx Runtime** `1.22.0`, **PyTorch** [`2.8.0a0+5228986c39.nv25.6`](https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform-release-notes/pytorch-jetson-rel.html), **Python** `3.12` and as well as _ensembles_. * ONNX Runtime backend does not support the OpenVINO and TensorRT execution providers. The CUDA execution provider is in Beta. * System shared memory is supported on Jetson. CUDA shared memory is not supported. * GPU metrics, GCS storage, S3 storage and Azure storage are not supported. The tar file contains the Triton server executable and shared libraries and also the C++ and Python client libraries and examples. For more information on how to install and use Triton on JetPack refer to [`jetson.md`](https://github.com/triton-inference-server/server/blob/r25.06/docs/user_guide/jetson.md). The wheel for the Python client library is present in the tar file and can be installed by running the following command: ``` python3 -m pip install --upgrade clients/python/tritonclient-2.59.0-py3-none-manylinux2014_aarch64.whl[all] ``` </details> <details> <summary><h2>Triton TRT-LLM Container Support Matrix</h2></summary> The Triton TensorRT-LLM container is built from the 25.04 image [`nvcr.io/nvidia/tritonserver:25.04-py3-min`](http://nvcr.io/nvidia/tritonserver:25.04-py3-min). Please refer to the [support matrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html) and [compatibility.md](https://github.com/triton-inference-server/server/blob/v2.59.0/docs/introduction/compatibility.md#container-name-trtllm-python-py3) for all dependency versions related to 25.04. However, the packages listed below have different versions than those specified in the support matrix. | Dependency | Version | | :------------: | :---------------: | | TensorRT-LLM | 0.20.0 | | TensorRT | 10.10.0.31 | </details>

<!-- TPRD-1381 --> # Triton Inference Server The Triton Inference Server provides a cloud inferencing solution optimized for both CPUs and GPUs. The server provides an inference service via an HTTP or GRPC endpoint, allowing remote clients to request inferencing for any model being managed by the server. For edge deployments, Triton Server is also available as a shared library with an API that allows the full functionality of the server to be included directly in an application. <details> <summary><h2>New Features and Improvements</h2></summary> - Exposed gRPC infer thread count as a server option. - Improved server stability during the gRPC client cancellation. - Improved server stability in tracing mode. - Added BLS decoupled request cancellation in the Python Backend - GenAI-Perf now offers a new configuration file alongside the command line. - GenAI-Perf now supports the Huggingface TGI generated endpoint. - GenAI-Perf added a Token per second per user (TPS/user) metric. - GenAI-Perf metric parsing speed was increased by 60%. </details> <details> <summary><h2>Security</h2></summary> vLLM containers include vllm version 0.8.1 which is affected by new vulnerabilities. Workarounds: Prior to the fix, your options include: - Do not expose the vLLM host to a network where any untrusted connections may reach the host. - Ensure that only the other vLLM hosts are able to connect to the TCP port used for the XPUB socket. Note that port used is random. </details> <details> <summary><h2>Known Issues</h2></summary> * The core Python binding may incur an additional D2H and H2D copy if the backend and frontend both specify device memory to be used for response tensors. * A segmentation fault related to DCGM and NSCQ may be encountered during server shutdown on NVSwitch systems. A possible workaround for this issue is to disable the collection of GPU metrics `tritonserver --allow-gpu-metrics false ...` * vLLM backend currently does not take advantage of the [vLLM v0.6](https://blog.vllm.ai/2024/09/05/perf-update.html) performance improvement when metrics are enabled. * Incorrect results are known to occur when using TensorRT (TRT) Backend for inference using int8 data type for I/O on the Blackwell GPU architecture. * When using TensorRT models, if auto-complete configuration is disabled and `is_non_linear_format_io:true` for [reformat-free tensors](https://github.com/triton-inference-server/server/blob/r24.08/docs/user_guide/model_configuration.md#non-linear-io-formats) is not provided in the model configuration, the model may not load successfully. * When using Python models in [decoupled mode](https://github.com/triton-inference-server/python_backend/tree/main?tab=readme-ov-file#decoupled-mode), users need to ensure that the `ResponseSender` goes out of scope or is properly cleaned up before unloading the model to guarantee that the unloading process executes correctly. * Restart support was temporarily removed for Python models. * Triton Inference Server with vLLM backend currently does not support running vLLM models with tensor parallelism sizes greater than 1 and the default "distributed_executor_backend" setting when using explicit model control mode. In attempt to load a vllm model (tp > 1) in explicit mode, users could potentially see failure at `initialize` step: `could not acquire lock for <_io.BufferedWriter name='<stdout>'> at interpreter shutdown, possibly due to daemon threads`. For the default model control mode, after server shutdown, vllm related sub-processes are not killed. Related vllm issue: https://github.com/vllm-project/vllm/issues/6766 . Please specify "distributed_executor_backend":"ray" in the `model.json` when deploying vllm models with tensor parallelism > 1. * When loading models with file override, multiple model configuration files are not supported. Users must provide the model configuration by setting parameter `"config" : "<JSON>"` instead of custom configuration file in the following format: `"file:configs/<model-config-name>.pbtxt" : "<base64-encoded-file-content>"`. * TensorRT-LLM [backend](https://github.com/triton-inference-server/tensorrtllm_backend) provides limited support of Triton extensions and features. * The TensorRT-LLM backend may core dump on server shutdown. This impacts server teardown only and will not impact inferencing. * The Java CAPI is known to have intermittent segfaults. * Some systems which implement `malloc()` may not release memory back to the operating system right away causing a false memory leak. This can be mitigated by using a different malloc implementation. `TCMalloc` and `jemalloc` are installed in the Triton container and can be [used by specifying the library in LD_PRELOAD](https://github.com/triton-inference-server/server/blob/r25.01/docs/user_guide/model_management.md). NVIDIA recommends experimenting with both `tcmalloc` and `jemalloc` to determine which one works better for your use case. * Auto-complete may cause an increase in server start time. To avoid a start time increase, users can provide the full model configuration and launch the server with `--disable-auto-complete-config`. * Auto-complete does not support PyTorch models due to lack of metadata in the model. It can only verify that the number of inputs and the input names matches what is specified in the model configuration. There is no model metadata about the number of outputs and datatypes. Related PyTorch bug:[ https://github.com/pytorch/pytorch/issues/38273](https://github.com/pytorch/pytorch/issues/38273) * Triton Client PIP wheels for ARM SBSA are not available from PyPI and pip will install an incorrect Jetson version of Triton Client library for Arm SBSA. The correct client wheel file can be pulled directly from the Arm SBSA SDK image and manually installed. * Traced models in PyTorch seem to create overflows when int8 tensor values are transformed to int32 on the GPU. Refer to [pytorch/pytorch#66930](https://github.com/pytorch/pytorch/issues/66930) for more information. * Triton cannot retrieve GPU metrics with [MIG-enabled GPU devices](https://docs.nvidia.com/datacenter/tesla/mig-user-guide/index.html#supported-gpus). * Triton metrics might not work if the host machine is running a separate DCGM agent on bare-metal or in a container. * When cloud storage (AWS, GCS, AZURE) is used as a model repository and a model has multiple versions, Triton creates an extra local copy of the cloud model’s folder in the temporary directory, which is deleted upon server’s shutdown. * Python backend support for Windows is limited and does not currently support the following features: - GPU tensors - CPU and GPU-related metrics - Custom execution environments - The model load/unload APIs </details> <details> <summary><h2>Client Libraries and Examples</h2></summary> Ubuntu 24.04 builds of the client libraries and examples are included in this release in the attached `v2.57.0_ubuntu2404.clients.tar.gz` file. The SDK is also available for as an Ubuntu 24.04 based [NGC Container](https://ngc.nvidia.com/catalog/containers/nvidia:tritonserver/tags). The SDK container includes the client libraries and examples, Performance Analyzer and Model Analyzer. Some components are also available in the tritonclient pip package. See [Getting the Client Libraries](https://github.com/triton-inference-server/client/tree/r25.04#getting-the-client-libraries-and-examples) for more information on each of these options. </details> <details> <summary><h2>Windows Support</h2></summary> > [!NOTE] > There is no Windows release for 25.04, the latest release is [25.02](https://github.com/triton-inference-server/server/releases/tag/v2.55.0). </details> <details> <summary><h2>Jetson iGPU Support</h2></summary> A release of Triton for [IGX](https://www.nvidia.com/en-us/edge-computing/products/igx/) is provided in the attached tar file: [`tritonserver2.56.0-igpu.tgz`](https://github.com/triton-inference-server/server/releases/download/v2.56.0/tritonserver2.56.0-igpu.tgz). * This release supports **TensorRT** `10.9.0.34`, **Onnx Runtime** `1.21.0`, **PyTorch** [`2.7.0a0+79aa17489c.nv25.4`](https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform-release-notes/pytorch-jetson-rel.html), **Python** `3.12` and as well as _ensembles_. * ONNX Runtime backend does not support the OpenVINO and TensorRT execution providers. The CUDA execution provider is in Beta. * System shared memory is supported on Jetson. CUDA shared memory is not supported. * GPU metrics, GCS storage, S3 storage and Azure storage are not supported. The tar file contains the Triton server executable and shared libraries and also the C++ and Python client libraries and examples. For more information on how to install and use Triton on JetPack refer to [`jetson.md`](https://github.com/triton-inference-server/server/blob/r25.04/docs/user_guide/jetson.md). The wheel for the Python client library is present in the tar file and can be installed by running the following command: ``` python3 -m pip install --upgrade clients/python/tritonclient-2.57.0-py3-none-manylinux2014_aarch64.whl[all] ``` </details> <details> <summary><h2>Triton TRT-LLM Container Support Matrix</h2></summary> The Triton TensorRT-LLM container is built from the 25.03 image [`nvcr.io/nvidia/tritonserver:25.03-py3-min`](http://nvcr.io/nvidia/tritonserver:25.03-py3-min). Please refer to the [support matrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html) for all dependency versions related to 25.03. However, the packages listed below have different versions than those specified in the support matrix. | Dependency | Version | | :------------: | :---------------: | | TensorRT-LLM | 0.18.2 | | TensorRT | 10.9.0.34 | </details>

<!-- TPRD-1195 --> # Triton Inference Server The Triton Inference Server provides a cloud inferencing solution optimized for both CPUs and GPUs. The server provides an inference service via an HTTP or GRPC endpoint, allowing remote clients to request inferencing for any model being managed by the server. For edge deployments, Triton Server is also available as a shared library with an API that allows the full functionality of the server to be included directly in an application. <details> <summary><h2>New Features and Improvements</h2></summary> * The Tensorflow Backend has been deprecated starting in 25.03. The last release of Triton Inference Server with the Tensorflow Backend is 25.02. Users wishing to continue using the Tensorflow Backend in 25.03 and later can build the [Tensorflow Backend](https://github.com/triton-inference-server/tensorflow_backend?tab=readme-ov-file#build-the-tensorflow-backend) from source and install the result into the `/opt/tritonserver/backends/` directory. * The “XX.YY-tf2-python-py3” container will no longer be available starting in 25.03. See the Tensorflow Backend deprecation. * Added generate and generate_stream inference types to SageMaker server. Customers can choose which inference types - infer (default), generate or generate_stream using SAGEMAKER_TRITON_INFERENCE_TYPE environment variable during server launch. * In an effort to allow quick, on-demand metric retrieval for external load balancers such as the [Kubernetes Inference Gateway API](https://gateway-api-inference-extension.sigs.k8s.io/), Triton when used with TRT-LLM can include live KV-cache utilization and capacity metrics in the HTTP response header when processing inference requests. </details> <details> <summary><h2>Known Issues</h2></summary> * The core Python binding may incur an additional D2H and H2D copy if the backend and frontend both specify device memory to be used for response tensors. * A segmentation fault related to DCGM and NSCQ may be encountered during server shutdown on NVSwitch systems. A possible workaround for this issue is to disable the collection of GPU metrics `tritonserver --allow-gpu-metrics false ...` * vLLM backend currently does not take advantage of the [vLLM v0.6](https://blog.vllm.ai/2024/09/05/perf-update.html) performance improvement when metrics are enabled. * Incorrect results are known to occur when using TensorRT (TRT) Backend for inference using int8 data type for I/O on the Blackwell GPU architecture. * When using TensorRT models, if auto-complete configuration is disabled and `is_non_linear_format_io:true` for [reformat-free tensors](https://github.com/triton-inference-server/server/blob/r24.08/docs/user_guide/model_configuration.md#non-linear-io-formats) is not provided in the model configuration, the model may not load successfully. * When using Python models in [decoupled mode](https://github.com/triton-inference-server/python_backend/tree/main?tab=readme-ov-file#decoupled-mode), users need to ensure that the `ResponseSender` goes out of scope or is properly cleaned up before unloading the model to guarantee that the unloading process executes correctly. * Restart support was temporarily removed for Python models. * Triton Inference Server with vLLM backend currently does not support running vLLM models with tensor parallelism sizes greater than 1 and the default "distributed_executor_backend" setting when using explicit model control mode. In attempt to load a vllm model (tp > 1) in explicit mode, users could potentially see failure at `initialize` step: `could not acquire lock for <_io.BufferedWriter name='<stdout>'> at interpreter shutdown, possibly due to daemon threads`. For the default model control mode, after server shutdown, vllm related sub-processes are not killed. Related vllm issue: https://github.com/vllm-project/vllm/issues/6766 . Please specify "distributed_executor_backend":"ray" in the `model.json` when deploying vllm models with tensor parallelism > 1. * When loading models with file override, multiple model configuration files are not supported. Users must provide the model configuration by setting parameter `"config" : "<JSON>"` instead of custom configuration file in the following format: `"file:configs/<model-config-name>.pbtxt" : "<base64-encoded-file-content>"`. * TensorRT-LLM [backend](https://github.com/triton-inference-server/tensorrtllm_backend) provides limited support of Triton extensions and features. * The TensorRT-LLM backend may core dump on server shutdown. This impacts server teardown only and will not impact inferencing. * The Java CAPI is known to have intermittent segfaults. * Some systems which implement `malloc()` may not release memory back to the operating system right away causing a false memory leak. This can be mitigated by using a different malloc implementation. `TCMalloc` and `jemalloc` are installed in the Triton container and can be [used by specifying the library in LD_PRELOAD](https://github.com/triton-inference-server/server/blob/r25.01/docs/user_guide/model_management.md). NVIDIA recommends experimenting with both `tcmalloc` and `jemalloc` to determine which one works better for your use case. * Auto-complete may cause an increase in server start time. To avoid a start time increase, users can provide the full model configuration and launch the server with `--disable-auto-complete-config`. * Auto-complete does not support PyTorch models due to lack of metadata in the model. It can only verify that the number of inputs and the input names matches what is specified in the model configuration. There is no model metadata about the number of outputs and datatypes. Related PyTorch bug:[ https://github.com/pytorch/pytorch/issues/38273](https://github.com/pytorch/pytorch/issues/38273) * Triton Client PIP wheels for ARM SBSA are not available from PyPI and pip will install an incorrect Jetson version of Triton Client library for Arm SBSA. The correct client wheel file can be pulled directly from the Arm SBSA SDK image and manually installed. * Traced models in PyTorch seem to create overflows when int8 tensor values are transformed to int32 on the GPU. Refer to [pytorch/pytorch#66930](https://github.com/pytorch/pytorch/issues/66930) for more information. * Triton cannot retrieve GPU metrics with [MIG-enabled GPU devices](https://docs.nvidia.com/datacenter/tesla/mig-user-guide/index.html#supported-gpus). * Triton metrics might not work if the host machine is running a separate DCGM agent on bare-metal or in a container. * When cloud storage (AWS, GCS, AZURE) is used as a model repository and a model has multiple versions, Triton creates an extra local copy of the cloud model’s folder in the temporary directory, which is deleted upon server’s shutdown. * Python backend support for Windows is limited and does not currently support the following features: - GPU tensors - CPU and GPU-related metrics - Custom execution environments - The model load/unload APIs </details> <details> <summary><h2>Client Libraries and Examples</h2></summary> Ubuntu 24.04 builds of the client libraries and examples are included in this release in the attached `v2.56.0_ubuntu2404.clients.tar.gz` file. The SDK is also available for as an Ubuntu 24.04 based [NGC Container](https://ngc.nvidia.com/catalog/containers/nvidia:tritonserver/tags). The SDK container includes the client libraries and examples, Performance Analyzer and Model Analyzer. Some components are also available in the tritonclient pip package. See [Getting the Client Libraries](https://github.com/triton-inference-server/client/tree/r25.03#getting-the-client-libraries-and-examples) for more information on each of these options. </details> <details> <summary><h2>Windows Support</h2></summary> > [!NOTE] > There is no Windows release for 25.03, the latest release is [25.02](https://github.com/triton-inference-server/server/releases/tag/v2.55.0). </details> <details> <summary><h2>Jetson iGPU Support</h2></summary> A release of Triton for [IGX](https://www.nvidia.com/en-us/edge-computing/products/igx/) is provided in the attached tar file: [`tritonserver2.56.0-igpu.tgz`](https://github.com/triton-inference-server/server/releases/download/v2.56.0/tritonserver2.56.0-igpu.tgz). * This release supports **TensorRT** `10.9.0.34`, **Onnx Runtime** `1.21.0`, **PyTorch** [`2.7.0a0+7c8ec84dab.nv25.3`](https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform-release-notes/pytorch-jetson-rel.html), **Python** `3.12` and as well as _ensembles_. * ONNX Runtime backend does not support the OpenVINO and TensorRT execution providers. The CUDA execution provider is in Beta. * System shared memory is supported on Jetson. CUDA shared memory is not supported. * GPU metrics, GCS storage, S3 storage and Azure storage are not supported. The tar file contains the Triton server executable and shared libraries and also the C++ and Python client libraries and examples. For more information on how to install and use Triton on JetPack refer to [`jetson.md`](https://github.com/triton-inference-server/server/blob/r25.03/docs/user_guide/jetson.md). The wheel for the Python client library is present in the tar file and can be installed by running the following command: ``` python3 -m pip install --upgrade clients/python/tritonclient-2.56.0-py3-none-manylinux2014_aarch64.whl[all] ``` </details> <details> <summary><h2>Triton TRT-LLM Container Support Matrix</h2></summary> The Triton TensorRT-LLM container is built from the 25.03 image [`nvcr.io/nvidia/tritonserver:25.03-py3-min`](http://nvcr.io/nvidia/tritonserver:25.03-py3-min). Please refer to the [support matrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html) for all dependency versions related to 25.03. However, the packages listed below have different versions than those specified in the support matrix. | Dependency | Version | | :------------: | :---------------: | | TensorRT-LLM | 0.18.0 | | TensorRT | 10.9.0.34 | </details>

# Triton Inference Server The Triton Inference Server provides a cloud inferencing solution optimized for both CPUs and GPUs. The server provides an inference service via an HTTP or GRPC endpoint, allowing remote clients to request inferencing for any model being managed by the server. For edge deployments, Triton Server is also available as a shared library with an API that allows the full functionality of the server to be included directly in an application. <details> <summary><h2>New Features and Improvements</h2></summary> * Python backend now supports setting and retrieving [Inference Response Parameters](https://github.com/triton-inference-server/python_backend/tree/r25.02#inference-response-parameters) on InferenceResponse objects on model.py. * Optimized the core Python binding architecture leading to improved OpenAI frontend performance. * Added dynamic sampling parameter handling, improving flexibility and consistency across vllm interactions. Added support for “guided_generation” request parameter for efficient constrained decoding workflows. * Improved Multi-Lora handling in TRTLLM GRPC Client `end_to_end_grpc_client.py` * GenAI-Perf added the ability to format output using Jinja2 templates. * GenAI-Perf telemetry now supports multiple metric endpoints. * GenAI-Perf now supports increased corpus size, 90x the previously supported size. * GenAI-Perf now supports keys without values as input. * GenAI-Perf fixed the OSL issue due to Performance Analyzer not removing the first 4 bytes from output. * GenAI-Perf added a chat template option for the TRT-LLM engine. * Performance Analyzer fixed TRITON_ENABLE_GPU compile definition bug. * Performance Analyzer bumped minimum required C++ version to C++20. * Performance Analyzer modified to disallow user attempts to use concurrency and warmup with the schedule flag. </details> <details> <summary><h2>Known Issues</h2></summary> * The core Python binding may incur an additional D2H and H2D copy if the backend and frontend both specify device memory to be used for response tensors. * A segmentation fault related to DCGM and NSCQ may be encountered during server shutdown on NVSwitch systems. A possible workaround for this issue is to disable the collection of GPU metrics `tritonserver --allow-gpu-metrics false ...` * vLLM backend currently does not take advantage of the [vLLM v0.6](https://blog.vllm.ai/2024/09/05/perf-update.html) performance improvement when metrics are enabled. * Incorrect results are known to occur when using TensorRT (TRT) Backend for inference using int8 data type for I/O on the Blackwell GPU architecture. * When using TensorRT models, if auto-complete configuration is disabled and `is_non_linear_format_io:true` for [reformat-free tensors](https://github.com/triton-inference-server/server/blob/r24.08/docs/user_guide/model_configuration.md#non-linear-io-formats) is not provided in the model configuration, the model may not load successfully. * When using Python models in [decoupled mode](https://github.com/triton-inference-server/python_backend/tree/main?tab=readme-ov-file#decoupled-mode), users need to ensure that the `ResponseSender` goes out of scope or is properly cleaned up before unloading the model to guarantee that the unloading process executes correctly. * Restart support was temporarily removed for Python models. * Triton Inference Server with vLLM backend currently does not support running vLLM models with tensor parallelism sizes greater than 1 and the default "distributed_executor_backend" setting when using explicit model control mode. In attempt to load a vllm model (tp > 1) in explicit mode, users could potentially see failure at `initialize` step: `could not acquire lock for <_io.BufferedWriter name='<stdout>'> at interpreter shutdown, possibly due to daemon threads`. For the default model control mode, after server shutdown, vllm related sub-processes are not killed. Related vllm issue: https://github.com/vllm-project/vllm/issues/6766 . Please specify "distributed_executor_backend":"ray" in the `model.json` when deploying vllm models with tensor parallelism > 1. * When loading models with file override, multiple model configuration files are not supported. Users must provide the model configuration by setting parameter `"config" : "<JSON>"` instead of custom configuration file in the following format: `"file:configs/<model-config-name>.pbtxt" : "<base64-encoded-file-content>"`. * TensorRT-LLM [backend](https://github.com/triton-inference-server/tensorrtllm_backend) provides limited support of Triton extensions and features. * The TensorRT-LLM backend may core dump on server shutdown. This impacts server teardown only and will not impact inferencing. * The Java CAPI is known to have intermittent segfaults. * Some systems which implement `malloc()` may not release memory back to the operating system right away causing a false memory leak. This can be mitigated by using a different malloc implementation. `TCMalloc` and `jemalloc` are installed in the Triton container and can be [used by specifying the library in LD_PRELOAD](https://github.com/triton-inference-server/server/blob/r25.01/docs/user_guide/model_management.md). NVIDIA recommends experimenting with both `tcmalloc` and `jemalloc` to determine which one works better for your use case. * Auto-complete may cause an increase in server start time. To avoid a start time increase, users can provide the full model configuration and launch the server with `--disable-auto-complete-config`. * Auto-complete does not support PyTorch models due to lack of metadata in the model. It can only verify that the number of inputs and the input names matches what is specified in the model configuration. There is no model metadata about the number of outputs and datatypes. Related PyTorch bug:[ https://github.com/pytorch/pytorch/issues/38273](https://github.com/pytorch/pytorch/issues/38273) * Triton Client PIP wheels for ARM SBSA are not available from PyPI and pip will install an incorrect Jetson version of Triton Client library for Arm SBSA. The correct client wheel file can be pulled directly from the Arm SBSA SDK image and manually installed. * Traced models in PyTorch seem to create overflows when int8 tensor values are transformed to int32 on the GPU. Refer to [pytorch/pytorch#66930](https://github.com/pytorch/pytorch/issues/66930) for more information. * Triton cannot retrieve GPU metrics with [MIG-enabled GPU devices](https://docs.nvidia.com/datacenter/tesla/mig-user-guide/index.html#supported-gpus). * Triton metrics might not work if the host machine is running a separate DCGM agent on bare-metal or in a container. * When cloud storage (AWS, GCS, AZURE) is used as a model repository and a model has multiple versions, Triton creates an extra local copy of the cloud model’s folder in the temporary directory, which is deleted upon server’s shutdown. * Python backend support for Windows is limited and does not currently support the following features: - GPU tensors - CPU and GPU-related metrics - Custom execution environments - The model load/unload APIs </details> <details> <summary><h2>Client Libraries and Examples</h2></summary> Ubuntu 24.04 builds of the client libraries and examples are included in this release in the attached `v2.55.0_ubuntu2404.clients.tar.gz` file. The SDK is also available for as an Ubuntu 24.04 based [NGC Container](https://ngc.nvidia.com/catalog/containers/nvidia:tritonserver/tags). The SDK container includes the client libraries and examples, Performance Analyzer and Model Analyzer. Some components are also available in the tritonclient pip package. See [Getting the Client Libraries](https://github.com/triton-inference-server/client/tree/r25.02#getting-the-client-libraries-and-examples) for more information on each of these options. </details> <details> <summary><h2>Windows Support</h2></summary> > [!NOTE] > The 25.02 Windows release is under development. </details> <details> <summary><h2>Jetson iGPU Support</h2></summary> A release of Triton for [IGX](https://www.nvidia.com/en-us/edge-computing/products/igx/) is provided in the attached tar file: [`tritonserver2.55.0-igpu.tgz`](https://github.com/triton-inference-server/server/releases/download/v2.55.0/tritonserver2.55.0-igpu.tgz). * This release supports **TensorFlow** `2.17.0`, **TensorRT** `10.8.0.40`, **Onnx Runtime** `1.20.1`, **PyTorch** [`2.6.0a0+ecf3bae40a`](https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform-release-notes/pytorch-jetson-rel.html), **Python** `3.12` and as well as _ensembles_. * ONNX Runtime backend does not support the OpenVINO and TensorRT execution providers. The CUDA execution provider is in Beta. * System shared memory is supported on Jetson. CUDA shared memory is not supported. * GPU metrics, GCS storage, S3 storage and Azure storage are not supported. The tar file contains the Triton server executable and shared libraries and also the C++ and Python client libraries and examples. For more information on how to install and use Triton on JetPack refer to [`jetson.md`](https://github.com/triton-inference-server/server/blob/r25.02/docs/user_guide/jetson.md). The wheel for the Python client library is present in the tar file and can be installed by running the following command: ``` python3 -m pip install --upgrade clients/python/tritonclient-2.55.0-py3-none-manylinux2014_aarch64.whl[all] ``` </details> <details> <summary><h2>Triton TRT-LLM Container Support Matrix</h2></summary> The Triton TensorRT-LLM container is built from the 25.02 image [`nvcr.io/nvidia/tritonserver:25.02-py3-min`](http://nvcr.io/nvidia/tritonserver:25.02-py3-min). Please refer to the [support matrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html) for all dependency versions related to 25.02. However, the packages listed below have different versions than those specified in the support matrix. | Dependency | Version | | :------------: | :---------------: | | TensorRT-LLM | 0.17.0 | | TensorRT | 10.8.0.40 | </details>

# Triton Inference Server The Triton Inference Server provides a cloud inferencing solution optimized for both CPUs and GPUs. The server provides an inference service via an HTTP or GRPC endpoint, allowing remote clients to request inferencing for any model being managed by the server. For edge deployments, Triton Server is also available as a shared library with an API that allows the full functionality of the server to be included directly in an application. <details> <summary><h2>New Features and Improvements</h2></summary> * Starting with the 25.01 release, Triton Inference Server supports Blackwell GPU architectures. * Starting from 25.01, the vLLM container shipped by Triton is NVIDIA optimized. Users who wish to use the public version of vLLM can continue to build a Triton-vLLM container on their end. * Fixed a bug when passing the correlation ID of string type to python_backend. Added datatype checks to correlation ID values. * vLLM backend can now take advantage of the vLLM v0.6 performance improvement by communicating with the vLLM engine via ZMQ. * GenAI-Perf now provides the exact input sequence length requested for synthetic text generation * GenAI-Perf supports the creation of a prefix pool to emulate system prompts via `--num-system-prompts` and `--system-prompt-length` * GenAI-Perf improved error visibility via returning more detailed errors when OpenAI frontends return an error or metric generation fails * GenAI-Perf reports time-to-second-token and request count in its metrics * GenAI-Perf allows the use of a custom tokenizer in its “compare” subcommand for comparing multiple profiles * GenAI-Perf natively supports `--request-count` for sending a specific number of requests and `--header` for sending a list of headers with every request * Model Analyzer functionality has been migrated to GenAI-Perf via the “analyze” subcommand, enabling the tool to sweep and find the optimal model configuration * A bytes appending bug was fixed in GenAI-Perf, resulting in more accurate output sequence lengths for Triton </details> <details> <summary><h2>Known Issues</h2></summary> * A segmentation fault related to DCGM and NSCQ may be encountered during server shutdown on NVSwitch systems. A possible workaround for this issue is to disable the collection of GPU metrics `tritonserver --allow-gpu-metrics false ...` * vLLM backend currently does not take advantage of the [vLLM v0.6](https://blog.vllm.ai/2024/09/05/perf-update.html) performance improvement when metrics are enabled. * Please note, that the vllm version provided in 25.01 container is 0.6.3.post1. Due to some issues with vllm library versioning, `vllm.__version__` displays `0.6.3`. * Incorrect results are known to occur when using TensorRT (TRT) Backend for inference using int8 data type for I/O on the Blackwell GPU architecture. * When running Torch TRT models, the output may differ from running the same model on a previous release. * When using TensorRT models, if auto-complete configuration is disabled and `is_non_linear_format_io:true` for [reformat-free tensors](https://github.com/triton-inference-server/server/blob/r24.08/docs/user_guide/model_configuration.md#non-linear-io-formats) is not provided in the model configuration, the model may not load successfully. * When using Python models in [decoupled mode](https://github.com/triton-inference-server/python_backend/tree/main?tab=readme-ov-file#decoupled-mode), users need to ensure that the `ResponseSender` goes out of scope or is properly cleaned up before unloading the model to guarantee that the unloading process executes correctly. * Restart support was temporarily removed for Python models. * Triton Inference Server with vLLM backend currently does not support running vLLM models with tensor parallelism sizes greater than 1 and the default "distributed_executor_backend" setting when using explicit model control mode. In attempt to load a vllm model (tp > 1) in explicit mode, users could potentially see failure at `initialize` step: `could not acquire lock for <_io.BufferedWriter name='<stdout>'> at interpreter shutdown, possibly due to daemon threads`. For the default model control mode, after server shutdown, vllm related sub-processes are not killed. Related vllm issue: https://github.com/vllm-project/vllm/issues/6766 . Please specify "distributed_executor_backend":"ray" in the `model.json` when deploying vllm models with tensor parallelism > 1. * When loading models with file override, multiple model configuration files are not supported. Users must provide the model configuration by setting parameter `"config" : "<JSON>"` instead of custom configuration file in the following format: `"file:configs/<model-config-name>.pbtxt" : "<base64-encoded-file-content>"`. * TensorRT-LLM [backend](https://github.com/triton-inference-server/tensorrtllm_backend) provides limited support of Triton extensions and features. * The TensorRT-LLM backend may core dump on server shutdown. This impacts server teardown only and will not impact inferencing. * The Java CAPI is known to have intermittent segfaults. * Some systems which implement `malloc()` may not release memory back to the operating system right away causing a false memory leak. This can be mitigated by using a different malloc implementation. `TCMalloc` and `jemalloc` are installed in the Triton container and can be [used by specifying the library in LD_PRELOAD](https://github.com/triton-inference-server/server/blob/r25.01/docs/user_guide/model_management.md). NVIDIA recommends experimenting with both `tcmalloc` and `jemalloc` to determine which one works better for your use case. * Auto-complete may cause an increase in server start time. To avoid a start time increase, users can provide the full model configuration and launch the server with `--disable-auto-complete-config`. * Auto-complete does not support PyTorch models due to lack of metadata in the model. It can only verify that the number of inputs and the input names matches what is specified in the model configuration. There is no model metadata about the number of outputs and datatypes. Related PyTorch bug:[ https://github.com/pytorch/pytorch/issues/38273](https://github.com/pytorch/pytorch/issues/38273) * Triton Client PIP wheels for ARM SBSA are not available from PyPI and pip will install an incorrect Jetson version of Triton Client library for Arm SBSA. The correct client wheel file can be pulled directly from the Arm SBSA SDK image and manually installed. * Traced models in PyTorch seem to create overflows when int8 tensor values are transformed to int32 on the GPU. Refer to [pytorch/pytorch#66930](https://github.com/pytorch/pytorch/issues/66930) for more information. * Triton cannot retrieve GPU metrics with [MIG-enabled GPU devices](https://docs.nvidia.com/datacenter/tesla/mig-user-guide/index.html#supported-gpus). * Triton metrics might not work if the host machine is running a separate DCGM agent on bare-metal or in a container. * When cloud storage (AWS, GCS, AZURE) is used as a model repository and a model has multiple versions, Triton creates an extra local copy of the cloud model’s folder in the temporary directory, which is deleted upon server’s shutdown. * Python backend support for Windows is limited and does not currently support the following features: - GPU tensors - CPU and GPU-related metrics - Custom execution environments - The model load/unload APIs </details> <details> <summary><h2>Client Libraries and Examples</h2></summary> Ubuntu 24.04 builds of the client libraries and examples are included in this release in the attached `v2.54.0_ubuntu2404.clients.tar.gz` file. The SDK is also available for as an Ubuntu 24.04 based [NGC Container](https://ngc.nvidia.com/catalog/containers/nvidia:tritonserver/tags). The SDK container includes the client libraries and examples, Performance Analyzer and Model Analyzer. Some components are also available in the tritonclient pip package. See [Getting the Client Libraries](https://github.com/triton-inference-server/client/tree/r25.01#getting-the-client-libraries-and-examples) for more information on each of these options. </details> <details> <summary><h2>Windows Support</h2></summary> A beta release of Triton for Windows is provided in the attached file: `tritonserver2.54.9-win.zip`. This is a beta release so functionality is limited and performance is not optimized. Additional features and improved performance will be provided in future releases. Specifically in this release: * HTTP/REST and GRPC endpoints are supported. *ONNX models are supported by the ONNX Runtime backend. The ONNX Runtime version is `1.20.1`. The CPU, CUDA, and TensorRT execution providers are supported. The OpenVINO execution provider is not supported. * OpenVINO models are supported. The OpenVINO version is `2024.5.0`. * Prometheus metrics endpoint is not supported. * System and CUDA shared memory are not supported. Known Issues * In our internal testing, we observed large latency in retrieving inference results from HTTP client. We recommend using gRPC to circumvent this issue. To use the Windows version of Triton, you must install all the necessary dependencies on your Windows system. These dependencies are available in the [Dockerfile.win10.min](https://github.com/triton-inference-server/server/blob/r25.01/Dockerfile.win10.min). The Dockerfile includes the following CUDA-related components: * Python `3.12.3` * CUDA `12.6.3` * cuDNN `9.6.0.74` * TensorRT `10.7.0.23` </details> <details> <summary><h2>Jetson iGPU Support</h2></summary> A release of Triton for [IGX](https://www.nvidia.com/en-us/edge-computing/products/igx/) is provided in the attached tar file: [`tritonserver2.54.0-igpu.tgz`](https://github.com/triton-inference-server/server/releases/download/v2.54.0/tritonserver2.54.0-igpu.tgz). * This release supports **TensorFlow** `2.17.0`, **TensorRT** `10.8.0.40`, **Onnx Runtime** `1.20.1`, **PyTorch** [`2.6.0a0+ecf3bae40a`](https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform-release-notes/pytorch-jetson-rel.html), **Python** `3.12` and as well as _ensembles_. * ONNX Runtime backend does not support the OpenVINO and TensorRT execution providers. The CUDA execution provider is in Beta. * System shared memory is supported on Jetson. CUDA shared memory is not supported. * GPU metrics, GCS storage, S3 storage and Azure storage are not supported. The tar file contains the Triton server executable and shared libraries and also the C++ and Python client libraries and examples. For more information on how to install and use Triton on JetPack refer to [`jetson.md`](https://github.com/triton-inference-server/server/blob/r25.01/docs/user_guide/jetson.md). The wheel for the Python client library is present in the tar file and can be installed by running the following command: ``` python3 -m pip install --upgrade clients/python/tritonclient-2.54.0-py3-none-manylinux2014_aarch64.whl[all] ``` </details> <details> <summary><h2>Triton TRT-LLM Container Support Matrix</h2></summary> The Triton TensorRT-LLM container is built from the 25.01 image [`nvcr.io/nvidia/tritonserver:25.01-py3-min`](http://nvcr.io/nvidia/tritonserver:25.01-py3-min). Please refer to the [support matrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html) for all dependency versions related to 25.01. However, the packages listed below have different versions than those specified in the support matrix. | Dependency | Version | | :------------: | :---------------: | | TensorRT-LLM | 0.17.0 | | TensorRT | 10.8.0.43 | </details>