Don't add cast for TF batch norm when type isn't changing

apache · Jun 4, 2020 · ae2a5dc · ae2a5dc
1 parent 34c95a8
commit ae2a5dc
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/python/tvm/relay/frontend/tensorflow.py b/python/tvm/relay/frontend/tensorflow.py
@@ -1227,7 +1227,7 @@ def _impl(inputs, attr, params, mod):
             attr['data_format'] = attr['data_format'].decode("utf-8")
             if attr['data_format'] == 'NCHW':
                 axis = 1
-        if 'U' in attr:
+        if 'U' in attr and attr['U'].name != attr['T'].name:
             need_cast = True
             inputs[0] = _op.cast(inputs[0], dtype=attr['U'].name)
         # Check if mean and variance are empty