TensorRT 10.6 GA Parser Update

TensorRT 10.6 GA Release - 2024-11-5

For more details, see the 10.6 GA release notes

• Updated ONNX submodule version to 1.17.0
• Fix issue where conditional layers were incorrectly being added
• Updated local function metadata to contain more information
• Added support for parsing nodes with Quickly Deployable Plugins
• Fixed handling of optional outputs

TensorRT 10.5 GA Parser Update

TensorRT 10.5 GA Release - 2024-10-01

For more details, see the 10.5 GA release notes.

• Added support for real-valued STFT operations
• Improved error handling in IParser

TensorRT 10.4 GA Parser Update

For more details, see the 10.4 GA release notes.

Changes

• Added support for tensor axes for Pad operations
• Added support for BlackmanWindow, HammingWindow, and HannWindow operations
• Improved error handling in IParserRefitter
• Fixed kernel shape inference in multi-input convolutions

TensorRT 10.3 GA Parser Update

For more details, see the 10.3 GA Release Notes

Changes

• Added support for tensor axes inputs for Slice nodes
• Updated ScatterElements importer to use an updated plugin

TensorRT 10.2 GA Parser Update

For more details, see the 10.2 GA Release Notes.

Changed

• Improved error handling with new macros and classes
• Minor changes to op importers for GRU and Squeeze.

TensorRT 10.1 GA Parser Update

For more details, see the 10.1 GA Release Notes.

Added

• Added supportsModelV2 API
• Added support for DeformConv operation
• Added support for PluginV3 TensorRT Plugins

Changed

• Marked all IParser and IParserRefitter APIs as noexcept

TensorRT 10.0 GA Parser Update

For more details, see the 10.0 GA release notes.

Added

• Added support for building with with protobuf-lite
• Added support for empty inputs in custom plugin nodes

Changed

• Fixed issue when parsing and refitting models with nested BatchNormalization nodes

ONNX-TensorRT 8.6 GA Release

For more details, see the 8.6 GA release notes for new features added in TensorRT 8.6.

Added

• Added support for GroupNormalization, LayerNormalization,IsInf operations
• Added support for INT32 input types for Argmin, Argmax, and TopK
• Added support for ReverseSequence operators with dynamic shapes
• Added support for TopK operators with dynamic K values
• Added support for N-D Trilu
• Added OnnxParserFlag enum and setFlag interfaces to the ONNX parser to modify the default parsing behavior
• Added metadata tracking, now ONNX node metadata will be embedded into TensorRT layers

Changed

• All cast operations will now use the new CastLayer over the pervious IdentityLayer.
• Updated ONNX submodule to v1.13.1

ONNX-TensorRT 8.6 EA Release

For more details, see the 8.6 EA release notes for new features added in TensorRT 8.6

Added

• Added support for GroupNormalization, LayerNormalization, IsInf operations
• Added support for INT32 input types for Argmin, Argmax, and TopK
• Added support for ReverseSequence operators with dynamic shapes
• Added support for TopK operators with dynamic K values
• Added OnnxParserFlag enum and setFlag interfaces to the ONNX parser to modify the default parsing behavior
• Added metadata tracking, now ONNX node metadata will be embedded into TensorRT layers

Changed

• All cast operations will now use the new CastLayer over the pervious IdentityLayer.

ONNX-TensorRT 8.5 GA Release

For more details, see the 8.5 GA release notes for new features added in TensorRT 8.5

Added

• Added the RandomNormal, RandomUniform, MeanVarianceNormalization, RoiAlign, Mod, Trilu, GridSample and NonZero operations
• Added native support for the NonMaxSuppression operator
• Added support for importing ONNX networks with UINT8 I/O types

Fixed

• Fixed an issue with output padding with 1D deconv
• Fixed an issue when flattening 1D tensors
• Fixed an issue when parsing String attributes from TRT plugins
• Fixed an issue when importing If subgraphs with shared initializer names
• Fixied an issue when importing Loop subgraphs with INT_MAX trip counts

Removed

• Removed onnx2trt binary. See the README.md for alternative binaries to run ONNX model with TensorRT.