[`tf.debugging`](https://www.tensorflow.org/versions/master/api_docs/python/tf/debugging),
    [`tf.dtypes`](https://www.tensorflow.org/versions/master/api_docs/python/tf/dtypes),
    [`tf.image`](https://www.tensorflow.org/versions/master/api_docs/python/tf/image),
    [`tf.io`](https://www.tensorflow.org/versions/master/api_docs/python/tf/io),
    [`tf.linalg`](https://www.tensorflow.org/versions/master/api_docs/python/tf/linalg),

 - Insert a debug container in clusters that don’t have ephemeral containers enabled 
 - Modify a crashing container for easier debugging by changing its image, for example to busybox, or its command, for example, to `sleep 1d` so you have time to `kubectl exec`.
 - Inspect configuration files on a node's host filesystem

containing the Cholesky decompositions for all input submatrices `[..., :, :]`.

**Note**: The gradient computation on GPU is faster for large matrices but
not for large batch dimensions when the submatrices are small. In this
case it might be faster to use the CPU.
  }];

  let arguments = (ins