onnx to tensorRT error in documentation?

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

Dear Sir/Madam,

Could you check that the example in
https://docs.nvidia.com/deeplearning/sdk/tensorrt-developer-guide/index.html#python_topics
and in 3.2.5. Importing From ONNX Using Python
and in 2. Create the build, network, and parser:
is correct?

I got it working by (well, it does not crash immediately as before):

with trt.Builder(TRT_LOGGER) as builder, builder.create_network() as network, trt.OnnxParser(network, TRT_LOGGER) as parser:
    with open(model_path, 'rb') as model:
        parser.parse(model.read())

It would be also nice to have documentation how to to prediction with tensorRT once one has done this preparation…

Terveisin, Markus

2

Hi Markus,

Please see this ONNX Resnet50 example that comes with the TensorRT Release for a reference of both parsing and doing inference:

/usr/src/tensorrt/samples/python/introductory_parser_samples/onnx_resnet50.py