Imagenet cannot operate with my exported in ONNX-model. May be I done something wrong/ Could you help?

(Sorry for a question, I’m still a newbie in Jetson Nano DevKit and I cannot access Jupyter notebooks in DLI “Getting Started…”) Can someone please explain what I am doing wrong (Getting Started with AI on Jetson Nano) :

  1. I’ve created two classes of datasets in …/data/<collection_name>/labels.txt.
  2. I train them by → saved checkpoint to : …models/<collection_name>/checkpoint.pth.tar
  3. Then I export model by and model exported to: → …models/<collection_name>/resnet18.onnx
  4. And when I use with made before …models/<collection_name>/resnet-18.onnx (with label = …/data/<collection_name>/labels.txt)
    the camera opens screen without my classes recognitions. But imagenet try to operate with standard full classes set but without my classes
    (according text messages in camera view).
    In the console I can see the following messages:

    [TRT] device GPU, models/<collection_name>/resnet18.onnx initialized
    [TRT] loaded 4 class labels
    [TRT] didn’t load expexted number of class description (4 of 2)
    [TRT] imagenet – failed to load sysnet class description (4 / 4 of 2)
    [TRT] imagenet – failed to initialize.
    jetson.inference – imageNet failed to load built-in network ‘googlenet’