TensorRT Samples

Contents

1. "Hello World" Samples

Sample	Language	Format	Description
sampleOnnxMNIST	C++	ONNX	“Hello World” For TensorRT With ONNX
network_api_pytorch_mnist	Python	INetwork	“Hello World” For TensorRT Using Pytorch

2. TensorRT API Samples

Sample	Language	Format	Description
sampleAlgorithmSelector	C++	ONNX	Algorithm Selection API usage
sampleCharRNN	C++	INetwork	Building An RNN Network Layer By Layer
sampleDynamicReshape	C++	ONNX	Digit Recognition With Dynamic Shapes In TensorRT
sampleINT8API	C++	ONNX	Performing Inference In INT8 Precision
sampleNamedDimensions	C++	ONNX	Working with named input dimensions
sampleOnnxMnistCoordConvAC	C++	ONNX	Implementing CoordConv with a custom plugin
sampleIOFormats	C++	ONNX	Specifying TensorRT I/O Formats
trtexec	C++	All	TensorRT Command-Line Wrapper: trtexec
engine_refit_onnx_bidaf	Python	ONNX	refitting an engine built from an ONNX model via parsers.
introductory_parser_samples	Python	ONNX	Introduction To Importing Models Using TensorRT Parsers
onnx_packnet	Python	ONNX	TensorRT Inference Of ONNX Models With Custom Layers

3. Application Samples

Sample	Language	Format	Description
detectron2	Python	ONNX	Support for Detectron 2 Mask R-CNN R50-FPN 3x model in TensorRT
efficientdet	Python	ONNX	EfficientDet Object Detection with TensorRT
efficientnet	Python	ONNX	EfficientNet V1 and V2 Classification with TensorRT
tensorflow_object_detection_api	Python	ONNX	TensorFlow Object Detection API Models in TensorRT
yolov3_onnx	Python	ONNX	Object Detection Using YOLOv3 With TensorRT ONNX Backend

Known Limitations

• UFF converter and GraphSurgeon tools are only supported with Tensorflow 1.x
• For the UFF samples, please use the NVIDIA tf1 (Tensorflow 1.x) for running these tests or install Tensorflow 1.x manually.