Remove labels from audio #3513
Description
Most versions of the audio summary APIs only support writing audio data,
but the TB 1.x tensorboard.summary.audio ops also support attaching a
“label” (Markdown text) to each audio clip. This feature is not present
in any version of the TensorFlow summary APIs, including tf.contrib,
and is not present in the TB 2.x tensorboard.summary.audio API.
This feature was added in response to user request (#296), but we don’t
think that it has been widely adopted. The data can only be written by a
particular combination of writing APIs that is not commonly used, and
some standard Google-internal tools don’t support reading it.
The feature makes it harder to add generic data support to the audio
dashboard, so we’re removing it for now. If this proves unacceptable, we
can look into reinstating the functionality by a different mechanism,
likely a parallel tensor time series of labels only.