yolov3_onnx模型转换，官方samples记错

TensorRT-5.1.5.0/samples/python/yolov3_onnx  里面包含了基本的samples

pip 默认安装的requirements.txt 里面的环境 可能会有问题，所以还是自己搭建

我是在conda 环境下装的

一， yolov3_to_onnx 模型

python yolov3_to_onnx.py

Traceback (most recent call last):
  File "yolov3_to_onnx.py", line 812, in
    main()
  File "yolov3_to_onnx.py", line 805, in main
    onnx.checker.check_model(yolov3_model_def)
  File "/home/amax/miniconda3/envs/tensorrt_py27/lib/python2.7/site-packages/onnx/checker.py", line 82, in check_model
    C.check_model(model.SerializeToString())
onnx.onnx_cpp2py_export.checker.ValidationError: Input size 2 not in range [min=1, max=1].

==> Context: Bad node spec: input: "085_convolutional_lrelu" input: "086_upsample_scale" output: "086_upsample" name: "086_upsample" op_type: "Upsample" attribute { name: "mode" s: "nearest" type: STRING }

TensorRT-5.1.5.0/samples/python/yolov3_onnx的  转换模型的例子需要在pip install onnx=1.4.1， 环境超过1.5.0错错错！！为啥官方的例子自己都不更新，这是有多么不上心啊，用户还要去论坛找answer  ，unbelievable！

https://devtalk.nvidia.com/default/topic/1052153/jetson-nano/tensorrt-backend-for-onnx-on-jetson-nano/1

 

二，

TensorRT-5.1.5.0/python/     安装

pip install tensorrt-5.1.5.0-cp27-none-linux_x86_64.whl

pip install pycuda

conda install pillow

python  onnx_to_tensorrt.py

 

Downloading from https://github.com/pjreddie/darknet/raw/f86901f6177dfc6116360a13cc06ab680e0c86b0/data/dog.jpg, this may take a while...
100% [............................................................................] 163759 / 163759
Loading ONNX file from path yolov3.onnx...
Beginning ONNX file parsing
Completed parsing of ONNX file
Building an engine from file yolov3.onnx; this may take a while...