I have build a seq2seq model using nn.LSTMCELL in it. And I want to export a trained one to onnx form. However, when I do this, I have an error says:
UserWarning: ONNX export failed on ATen operator _thnn_fused_lstm_cell because torch.onnx.symbolic_opset9._thnn_fused_lstm_cell does not exist .format(op_name, opset_version, op_name))
I think this maybe due to the opset nort supporting nn.LSTMCELL.
Any one could help and share some advice?