diff --git a/onnx_coreml/_operators_nd.py b/onnx_coreml/_operators_nd.py
index 6bc6fdd..29b0fe0 100644
--- a/onnx_coreml/_operators_nd.py
+++ b/onnx_coreml/_operators_nd.py
@@ -40,7 +40,14 @@ def load_input_constants(builder, node, graph, err):
 def _add_conv_like_op(add_func, get_params_func, params_dict,
                       builder, node, graph, err):
 
+
+    # To do: Need to avoid dependence on rank for conversion since rank is not always available.
+
     rank = builder._get_rank(node.inputs[0])
+
+    if rank < 0 and node.op_type == "Conv" and "w_shape" in params_dict:
+        rank = len(params_dict["w_shape"])
+
     if rank == 4:
         get_params_func(builder, node, graph, err, params_dict)
         add_func(node.inputs, node.outputs, params_dict=params_dict, builder=builder, node=node, graph=graph, err=err)
@@ -67,7 +74,7 @@ def _add_conv_like_op(add_func, get_params_func, params_dict,
             axes=axes
         )
     else:
-        return err.unsupported_op_configuration(builder, node, graph, "provided number axes {} not supported".format(rank))
+        return err.unsupported_op_configuration(builder, node, graph, "Unable to infer rank of the input. Invalid rank {} not supported".format(rank))
 
     
 def add_broadcastable_op_chain(builder, node, err, add_op_function):