added comments and warning

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

@@ -55,11 +55,12 @@
 from __future__ import division
 from __future__ import print_function
 from __future__ import unicode_literals
+
 import re
+import logging
 import numpy as np
 from .export_onnx import MXNetGraph as mx_op
 
-
 def import_onnx_modules():
     """ To make sure ONNX is runtime dependency, it is imported used only when needed"""
     try:
@@ -94,6 +95,7 @@ def transform_padding(pad_width):
     onnx_pad_width = [0]*num_pad_values
 
     start_index = 0
+    # num_pad_values will always be multiple of 2
     end_index = int(num_pad_values/2)
     for idx in range(0, num_pad_values):
         if idx % 2 == 0:
@@ -552,6 +554,15 @@ def convert_pooling(node, **kwargs):
     input_node = proc_nodes[input_node_idx]
     name = node["name"]
 
+    pooling_convention = attrs.get('pooling_convention', 'valid')
+
+    if pooling_convention == 'full':
+        pooling_warning = "Pooling: ONNX currently doesn't support pooling_convention. " \
+                          "This might lead to shape or accuracy issues. " \
+                          "https://github.com/onnx/onnx/issues/549"
+
+        logging.warning(pooling_warning)
+
     pad_dims = list(parse_helper(attrs, "pad", [0, 0]))
     pad_dims = pad_dims + pad_dims
     pool_types = {"max": "MaxPool", "avg": "AveragePool"}