tf.assert_positive

assert_positive(x, data=None, summarize=None, message=None, name=None)

    Assert the condition `x > 0` holds element-wise.

    

    Example of adding a dependency to an operation:

    

    ```python

    with tf.control_dependencies([tf.assert_positive(x)]):

      output = tf.reduce_sum(x)

    ```

    

    Positive means, for every element `x[i]` of `x`, we have `x[i] > 0`.

    If `x` is empty this is trivially satisfied.

    

    Args:

      x:  Numeric `Tensor`.

      data:  The tensors to print out if the condition is False.  Defaults to

        error message and first few entries of `x`.

      summarize: Print this many entries of each tensor.

      message: A string to prefix to the default message.

      name: A name for this operation (optional).  Defaults to "assert_positive".

    

    Returns:

      Op raising `InvalidArgumentError` unless `x` is all positive.