I read the page regarding BYOM and followed the instruction by installing what’s necessary, but still encountered some problems.
The ONNX model I converted is a YOLOv8 model in pyTorch format.
The .pt file was downloaded from here:
I then tried to convert the said ONNX model using tao_byom but failed. The error message is about some keras things not being defined, but that’s another story. I now would like to know if it’s possible to convert a state-of-the-art model, i.e, YOLOv8, to a TAO-compatible model and then train it as I saw this:
"Bring Your Own Model (BYOM) is a Python-based package that converts any open-source ONNX model to a TAO-comaptible model. "
Does BYOM currently only support classification and segmentation tasks as they’re mentioned in the instruction article?