|View source on GitHub|
Write an audio summary.
tf.summary.audio( name, data, sample_rate, step=None, max_outputs=3, encoding=None, description=None )
name: A name for this summary. The summary tag used for TensorBoard will be this name prefixed by any active name scopes.
Tensorrepresenting audio data with shape
[k, t, c], where
kis the number of audio clips,
tis the number of frames, and
cis the number of channels. Elements should be floating-point values in
[-1.0, 1.0]. Any of the dimensions may be statically unknown (i.e.,
Tensorthat represents the sample rate, in Hz. Must be positive.
int64-castable monotonic step value for this summary. If omitted, this defaults to
tf.summary.experimental.get_step(), which must not be None.
intor rank-0 integer
Tensor. At most this many audio clips will be emitted at each step. When more than
max_outputsmany clips are provided, the first
max_outputsmany clips will be used and the rest silently discarded.
encoding: Optional constant
strfor the desired encoding. Only "wav" is currently supported, but this is not guaranteed to remain the default, so if you want "wav" in particular, set this explicitly.
description: Optional long-form description for this summary, as a constant
str. Markdown is supported. Defaults to empty.
True on success, or false if no summary was emitted because no default summary writer was available.
ValueError: if a default writer exists, but no step was provided and