Improve handling of symbolic dimensions in the onnxruntime_test.py script. (#3959)

If a symbolic dimension is found allow the user to provide a value, or default to 1.

`python .\onnxruntime_test.py --symbolic_dims batch=1,seqlen=4 onnxruntime\test\testdata\transform\fusion\fast_gelu_use_graph_input.onnx`

onnxruntime/python/tools/onnxruntime_test.py

@@ -31,6 +31,11 @@ def main():
     parser.add_argument('num_iters', nargs='?', type=int, default=1000, help='model run iterations. default=1000')
     parser.add_argument('--debug', action='store_true', help='pause execution to allow attaching a debugger.')
     parser.add_argument('--profile', action='store_true', help='enable chrome timeline trace profiling.')
+    parser.add_argument('--symbolic_dims', default=None, type=lambda s: dict(x.split("=") for x in s.split(",")),
+                        help='Comma separated name=value pairs for any symbolic dimensions in the model input. '
+                             'e.g. --symbolic_dims batch=1,seqlen=5. '
+                             'If not provided, the value of 1 will be used for all symbolic dimensions.')
+
     args = parser.parse_args()
     iters = args.num_iters
 
@@ -50,8 +55,21 @@ def main():
 
     feeds = {}
     for input_meta in sess.get_inputs():
-        # replace any symbolic dimensions (value is None) with 1
-        shape = [dim if dim else 1 for dim in input_meta.shape]
+        # replace any symbolic dimensions
+        shape = []
+        for dim in input_meta.shape:
+            if not dim:
+                # unknown dim
+                shape.append(1)
+            elif type(dim) == str:
+                # symbolic dim. see if we have a value otherwise use 1
+                if dim in args.symbolic_dims:
+                    shape.append(int(args.symbolic_dims[dim]))
+                else:
+                    shape.append(1)
+            else:
+                shape.append(dim)
+
         if input_meta.type in float_dict:
             feeds[input_meta.name] = np.random.rand(*shape).astype(float_dict[input_meta.type])
         elif input_meta.type in integer_dict: