Parses ONNX models for execution with TensorRT.
See also the TensorRT documentation.
Development on the Master branch is for the latest version of TensorRT 7.1 with full-dimensions and dynamic shape support.
For previous versions of TensorRT, refer to their respective branches.
Building INetwork objects in full dimensions mode with dynamic shape support requires calling the following API:
C++
const auto explicitBatch = 1U << static_cast<uint32_t>(nvinfer1::NetworkDefinitionCreationFlag::kEXPLICIT_BATCH);
builder->createNetworkV2(explicitBatch)
Python
import tensorrt
explicit_batch = 1 << (int)(tensorrt.NetworkDefinitionCreationFlag.EXPLICIT_BATCH)
builder.create_network(explicit_batch)
For examples of usage of these APIs see:
Current supported ONNX operators are found in the operator support matrix.
For building on master, we recommend following the instructions on the master branch of TensorRT to take advatange of the latest plugin code required for importing certain operators.
To build only the ONNX-TensorRT parser, follow the following steps:
cd onnx-tensorrt
mkdir build && cd build
cmake .. -DTENSORRT_ROOT=<path_to_trt> && make -j
// Ensure that you update your LD_LIBRARY_PATH to pick up the location of the newly built library:
export LD_LIBRARY_PATH=$PWD:$LD_LIBRARY_PATH
ONNX models can be converted to serialized TensorRT engines using the onnx2trt executable:
onnx2trt my_model.onnx -o my_engine.trt
ONNX models can also be converted to human-readable text:
onnx2trt my_model.onnx -t my_model.onnx.txt
ONNX models can also be optimized by ONNX's optimization libraries (added by dsandler).
To optimize an ONNX model and output a new one use -m to specify the output model name and -O to specify a semicolon-separated list of optimization passes to apply:
onnx2trt my_model.onnx -O "pass_1;pass_2;pass_3" -m my_model_optimized.onnx
See more all available optimization passes by running:
onnx2trt -p
See more usage information by running:
onnx2trt -h
Python bindings for the ONNX-TensorRT parser are packaged in the shipped .whl files. Install them with
pip install <tensorrt_install_dir>/python/tensorrt-7.x.x.x-cp27-none-linux_x86_64.whl
TensorRT 7.1 supports ONNX release 1.6.0. Install it with:
pip install onnx==1.6.0
The TensorRT backend for ONNX can be used in Python as follows:
import onnx
import onnx_tensorrt.backend as backend
import numpy as np
model = onnx.load("/path/to/model.onnx")
engine = backend.prepare(model, device='CUDA:1')
input_data = np.random.random(size=(32, 3, 224, 224)).astype(np.float32)
output_data = engine.run(input_data)[0]
print(output_data)
print(output_data.shape)The model parser library, libnvonnxparser.so, has its C++ API declared in this header:
NvOnnxParser.h
Build the onnx_tensorrt Docker image using tar-based TensorRT by running:
git clone --recurse-submodules https://github.com/onnx/onnx-tensorrt.git
cd onnx-tensorrt
cp /path/to/TensorRT-7.x.x.tar.gz .
docker build -f docker/onnx-tensorrt-tar.Dockerfile --tag=onnx-tensorrt:7.x.x .
Build the onnx_tensorrt Docker image using deb-based TensorRT by running:
git clone --recurse-submodules https://github.com/onnx/onnx-tensorrt.git
cd onnx-tensorrt
cp /path/to/nv-tensorrt-repo-ubuntu1x04-cudax.x-trt7.x.x.x-ga-yyyymmdd_1-1_amd64.deb .
docker build -f docker/onnx-tensorrt-deb.Dockerfile --tag=onnx-tensorrt:7.x.x.x .
After installation (or inside the Docker container), ONNX backend tests can be run as follows:
Real model tests only:
python onnx_backend_test.py OnnxBackendRealModelTest
All tests:
python onnx_backend_test.py
You can use -v flag to make output more verbose.
Pre-trained models in ONNX format can be found at the ONNX Model Zoo