Fix for issue apache#4831. The data_min_idx and data_max_idx were fli…

…pped. (apache#5136) Existing test cases cover this fix. In addition I have added an assert to make sure that the data_min is always less than equal to data_max.
neo-ai · Apr 17, 2020 · f26874c · f26874c
1 parent abeea08
commit f26874c
Showing 1 changed file with 3 additions and 2 deletions.
diff --git a/python/tvm/relay/frontend/mxnet.py b/python/tvm/relay/frontend/mxnet.py
@@ -1229,6 +1229,7 @@ def _get_data_scale_and_zp(_data, _inputs,
         """ Finds the Qnn params for the data expr. """
         data_min = _inputs[_data_min_idx]
         data_max = _inputs[_data_max_idx]
+        assert data_min <= data_max
         data_dtype = _infer_type(_data).checked_type.dtype
         assert data_dtype in {'int8', 'uint8'}
         if data_min < 0.0:
@@ -1420,8 +1421,8 @@ def _parse():
             ###############################################
             # Last 2 indexes are data min and max. If the conv has a sum, then last 2 indexes are
             # for the second tensor. So, the data min max indexes are last 3 and 4
-            data_min_idx = -1
-            data_max_idx = -2
+            data_min_idx = -2
+            data_max_idx = -1
             if has_sum:
                 data_min_idx = -4
                 data_max_idx = -3