在模型训练期间，出现错误“ ScanOp：预期的扫描轴在[-1，1）范围内，但得到1”，这似乎来自cumsum（）操作。我跟踪了代码，无法找出原因。有人可以帮忙吗？
下面是更详细的错误日志。
667.master | [2019-03-08T22:42:28Z] File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/math_ops.py", line 2373, in cumsum
667.master | [2019-03-08T22:42:28Z] x, axis, exclusive=exclusive, reverse=reverse, name=name)
667.master | [2019-03-08T22:42:28Z] File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/gen_math_ops.py", line 2340, in cumsum
667.master | [2019-03-08T22:42:28Z] name=name)
667.master | [2019-03-08T22:42:28Z] File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/op_def_library.py", line 787, in _apply_op_helper
667.master | [2019-03-08T22:42:28Z] op_def=op_def)
667.master | [2019-03-08T22:42:28Z] File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 3414, in create_op
667.master | [2019-03-08T22:42:28Z] op_def=op_def)
667.master | [2019-03-08T22:42:28Z] File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 1740, in __init__
667.master | [2019-03-08T22:42:28Z] self._traceback = self._graph._extract_stack() # pylint: disable=protected-access
667.master | [2019-03-08T22:42:28Z]
667.master | [2019-03-08T22:42:28Z] InvalidArgumentError (see above for traceback): ScanOp: Expected scan axis in the range [-1, 1), but got 1