Using NNVM customized model

#1

I am working with a frameworking that uses nnvm layers to be processed, and it doesn’t support relay yet.

There’s a tutorial that shows the way to define a network using NNVM. Here’s the part where they use a customized model. The full link is here: tune nnvm

 elif name == 'custom':
        # an example for custom network
        from nnvm.testing import utils
        net = nnvm.sym.Variable('data')
        net = nnvm.sym.conv2d(net, channels=4, kernel_size=(3,3), padding=(1,1))
        net = nnvm.sym.flatten(net)
        net = nnvm.sym.dense(net, units=1000)
        net, params = utils.create_workload(net, batch_size, (3, 224, 224))

However, It seems that the weights are random here. I am trying to load a the weights for the customized model. Is there a way to do that through NNVM. I couldn’t find any function in nnvm that can do that yet. For example in other frameworks like tensorflow, you can use something like "model.load_weights(‘model_weights.h5’)

Thank you very much.