Triton Inference Server Release 21.06.1

The Triton Inference Server container image, release 21.06.1, is available on NGC and is open source on GitHub.

Contents of the Triton Inference Server container

The Triton Inference Server Docker image contains the inference server executable and related shared libraries in /opt/tritonserver.

The container also includes the following:

Driver Requirements

Release 21.06.1 is based on NVIDIA CUDA 11.3.1, which requires NVIDIA Driver release 465.19.01 or later. However, if you are running on Data Center GPUs (formerly Tesla), for example, T4, you may use NVIDIA driver release 418.40 (or later R418), 440.33 (or later R440), 450.51 (or later R450), or 460.27 (or later R460). The CUDA driver's compatibility package only supports particular drivers. For a complete list of supported drivers, see the CUDA Application Compatibility topic. For more information, see CUDA Compatibility and Upgrades and NVIDIA CUDA and Drivers Support.

GPU Requirements

Release 21.06.1 supports CUDA compute capability 6.0 and higher. This corresponds to GPUs in the NVIDIA Pascal, Volta, Turing, and Ampere Architecture GPU families. Specifically, for a list of GPUs that this compute capability corresponds to, see CUDA GPUs. For additional support details, see Deep Learning Frameworks Support Matrix.

Key Features and Enhancements

This Inference Server release includes the following key features and enhancements.

  • The Forest Inference Library (FIL) backend is added to Triton. The FIL backend allows forest models trained by several popular machine learning frameworks (including XGBoost, LightGBM, Scikit-Learn, and cuML) to be deployed in a Triton.
  • Windows version of Triton now includes the OpenVino backend.
  • The Performance Analyzer (perf_analyzer) now supports testing against the Triton C API.
  • The Python backend now allows the use of conda to create a unique execution environment for your Python model. See https://github.com/triton-inference-server/python_backend#using-custom-python-execution-environments.
  • Python models that crash or exit unexpectedly are now automatically restarted by Triton.
  • Model repositories in S3 storage can now be accessed using HTTPS protocol. See https://github.com/triton-inference-server/server/blob/main/docs/user_guide/model_repository.md for more information.
  • Triton now collects GPU metrics for MIG partitions.
  • Passive model instances can now be specified in the model configuration. A passive model instance will be loaded and initialized by Triton, but no inference requests will be sent to the instance. Passive instances are typically used by a custom backend that uses its own mechanisms to distribute work to the passive instances. See the ModelInstanceGroup section of model_config.proto for the setting.
  • NVDLA support is added to the TensorRT backend.
  • ONNX Runtime version updated to 1.8.0.
  • Windows build documentation simplified and improved.
  • Improved detailed and summary reports in Model Analyzer.
  • Added an offline mode to Model Analyzer.
  • Refer to the 21.06 column of the Frameworks Support Matrix for container image versions that the 21.05 inference server container is based on.
  • Ubuntu 20.04 with May 2021 updates.

NVIDIA Triton Inference Server Container Versions

The following table shows what versions of Ubuntu, CUDA, Triton Inference Server, and TensorRT are supported in each of the NVIDIA containers for Triton Inference Server. For older container versions, refer to the Frameworks Support Matrix.

Known Issues

  • Triton’s TensorRT support depends on the input-consumed feature of TensorRT. In some rare cases using TensorRT 8.0 and earlier versions, the input-consumed event fires earlier than expected, causing Triton to overwrite input tensors while they are still in use and leading to corrupt input data being used for inference. This situation occurs when the inputs feed directly into a TensorRT layer that is optimized into a ForeignNode in the builder log. If you encounter accuracy issues with your TensorRT model, you can work around the issue by enabling the output_copy_stream option in your model’s configuration (https://github.com/triton-inference-server/common/blob/main/protobuf/model_config.proto#L816).
  • Running a PyTorch TorchScript model using the PyTorch backend, where multiple instances of a model are configured can lead to a slowdown in model execution due to the following PyTorch issue: https://github.com/pytorch/pytorch/issues/27902.
  • Compared with the 21.02 and earlier releases, there are backwards incompatible changes in the example Python client shared-memory support library when that library is used for tensors of type BYTES. The utils.serialize_byte_tensor() and utils.deserialize_byte_tensor() functions now return np.object_ numpy arrays where previously they returned np.bytes_ numpy arrays. Code depending on np.bytes_ must be updated. This change was necessary because the np.bytes_ type removes all trailing zeros from each array element and so binary sequences ending in zero(s) could not be represented with the old behavior. Correct usage of the Python client shared-memory support library is shown inhttps://github.com/triton-inference-server/server/blob/r21.03/src/clients/python/examples/simple_http_shm_string_client.py.