Remove labels from audio #3513
Assignees
Labels
Comments
wchargin
added a commit
that referenced
this issue
Apr 17, 2020
Summary: 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. It hasn’t been widely adopted, and doesn’t integrate nicely in its current form with upcoming efforts like migrating the audio dashboard to use the generic data APIs. This commit causes labels to no longer be read by TensorBoard. Labels are still written by the TB 1.x summary ops if specified, but now print a warning. Because we do not change the write paths, it is important that `audio.metadata` *not* set `DATA_CLASS_BLOB_SEQUENCE`, because otherwise audio summaries will not be touched by `dataclass_compat`. Tests for `dataclass_compat` cover this. We don’t think that this feature has wide adoption. If this change gets significant pushback, we can look into restoring labels with a different implementation, likely as a parallel tensor time series. Closes #3513. Test Plan: Unit tests updated. As a manual test, TensorBoard still works on both legacy audio data and audio data in the new form (that’s not actually written to disk yet), and simply does not display any labels. wchargin-branch: audio-no-labels
caisq
pushed a commit
to caisq/tensorboard
that referenced
this issue
May 19, 2020
Summary: 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. It hasn’t been widely adopted, and doesn’t integrate nicely in its current form with upcoming efforts like migrating the audio dashboard to use the generic data APIs. This commit causes labels to no longer be read by TensorBoard. Labels are still written by the TB 1.x summary ops if specified, but now print a warning. Because we do not change the write paths, it is important that `audio.metadata` *not* set `DATA_CLASS_BLOB_SEQUENCE`, because otherwise audio summaries will not be touched by `dataclass_compat`. Tests for `dataclass_compat` cover this. We don’t think that this feature has wide adoption. If this change gets significant pushback, we can look into restoring labels with a different implementation, likely as a parallel tensor time series. Closes tensorflow#3513. Test Plan: Unit tests updated. As a manual test, TensorBoard still works on both legacy audio data and audio data in the new form (that’s not actually written to disk yet), and simply does not display any labels. wchargin-branch: audio-no-labels
caisq
pushed a commit
that referenced
this issue
May 27, 2020
Summary: 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. It hasn’t been widely adopted, and doesn’t integrate nicely in its current form with upcoming efforts like migrating the audio dashboard to use the generic data APIs. This commit causes labels to no longer be read by TensorBoard. Labels are still written by the TB 1.x summary ops if specified, but now print a warning. Because we do not change the write paths, it is important that `audio.metadata` *not* set `DATA_CLASS_BLOB_SEQUENCE`, because otherwise audio summaries will not be touched by `dataclass_compat`. Tests for `dataclass_compat` cover this. We don’t think that this feature has wide adoption. If this change gets significant pushback, we can look into restoring labels with a different implementation, likely as a parallel tensor time series. Closes #3513. Test Plan: Unit tests updated. As a manual test, TensorBoard still works on both legacy audio data and audio data in the new form (that’s not actually written to disk yet), and simply does not display any labels. wchargin-branch: audio-no-labels
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Most versions of the audio summary APIs only support writing audio data,
but the TB 1.x
tensorboard.summary.audioops 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.audioAPI.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.
The text was updated successfully, but these errors were encountered: