...
There are two approaches which can be taken to import/export onnx model.
...
Through MXNet's symbolic operators
Implement at the MXNet layer by parsing the ONNX model(in protobuf format) and turn into MXNet Symbolic operators and build MXNet model directly. Similarly, MXNet model can be converted to ONNX format at this layer.
...
def import(..., format='onnx'):
# returns mxnet graph with parameters
...
return sym, params
def export(..., format='onnx'):
# returns ONNX protobuf object
...
return onnx_proto
...
Through
...
NNVM/top operators
The DMLC community has released the nnvm/tvm complier and an intermediate representation of the models.
...