Use inputs from get_inputs

python/mxnet/contrib/onnx/mx2onnx/_op_translations.py

@@ -226,15 +226,6 @@ def convert_deconvolution(node, **kwargs):
     """
     name, inputs, attrs = get_inputs(node, kwargs)
 
-    num_inputs = len(inputs)
-
-    proc_nodes = kwargs["proc_nodes"]
-    input_node = proc_nodes[kwargs["index_lookup"][inputs[0][0]]].name
-    weights_node = proc_nodes[kwargs["index_lookup"][inputs[1][0]]].name
-
-    if num_inputs > 2:
-        bias_node = proc_nodes[kwargs["index_lookup"][inputs[2][0]]].name
-
     kernel_dims = list(parse_helper(attrs, "kernel"))
     stride_dims = list(parse_helper(attrs, "stride", [1, 1]))
     pad_dims = list(parse_helper(attrs, "pad", [0, 0]))
@@ -244,13 +235,9 @@ def convert_deconvolution(node, **kwargs):
 
     pad_dims = pad_dims + pad_dims
 
-    input_nodes = [input_node, weights_node]
-    if num_inputs > 2:
-        input_nodes.append(bias_node)
-
     deconv_node = onnx.helper.make_node(
         "ConvTranspose",
-        inputs=input_nodes,
+        inputs=inputs,
         outputs=[name],
         kernel_shape=kernel_dims,
         strides=stride_dims,