How to export a pretrained model?

[Bresenham]

Is there a way to export a pretrained model to executorch? This example https://pytorch.org/executorch/stable/getting-started-setup.html#export-a-program only shows how to export a new model instance. I tried doing it like this

# 1. torch.export: Defines the program with the ATen operator set.
model.eval()
aten_dialect = torch.export.export( model, ( torch.ones( 2 ) ) )

# 2. to_edge: Make optimizations for Edge devices
edge_program = executorch.exir.to_edge( aten_dialect )

# 3. to_executorch: Convert the graph to an ExecuTorch program
executorch_program = edge_program.to_executorch()

# 4. Save the compiled .pte program
with open( "net.pte", "wb" ) as file:
    file.write(executorch_program.buffer)

but I get Expecting 'args' to be a tuple of example positional inputs, got <class 'torch.Tensor'>.

My model:

class Net( nn.Module ):
    def __init__( self ):
        super().__init__()
        self.inputFeatures = 2
        self.fc1 = nn.Linear(  self.inputFeatures,  1 )

    def forward( self, x ):
        fc1 = F.sigmoid( self.fc1( x ) )
        return fc1

[kimishpatel]

Maybe try with aten_dialect = torch.export.export( model, ( torch.ones( 2 ) , ) ). Notice the extra comma in examples tuple. I remember something of the effect should resolve it. If not, let us know

[dbort]

Please reopen if this didn't fix your issue, @Bresenham