Calculates the mean and variance of `x`. The mean and variance are calculated by aggregating the contents of `x` across `axes`. If `x` is 1-D and `axes = [0]` this is just the mean and variance of a vector. Note: shift is currently not used; the true mean is computed and used. When usi
(
x,
axes,
shift=None,
keepdims=False,
name=None)
| 1334 | |
| 1335 | @tf_export("nn.moments", v1=[]) |
| 1336 | def moments_v2( |
| 1337 | x, |
| 1338 | axes, |
| 1339 | shift=None, |
| 1340 | keepdims=False, |
| 1341 | name=None): |
| 1342 | """Calculates the mean and variance of `x`. |
| 1343 | |
| 1344 | The mean and variance are calculated by aggregating the contents of `x` |
| 1345 | across `axes`. If `x` is 1-D and `axes = [0]` this is just the mean |
| 1346 | and variance of a vector. |
| 1347 | |
| 1348 | Note: shift is currently not used; the true mean is computed and used. |
| 1349 | |
| 1350 | When using these moments for batch normalization (see |
| 1351 | `tf.nn.batch_normalization`): |
| 1352 | |
| 1353 | * for so-called "global normalization", used with convolutional filters with |
| 1354 | shape `[batch, height, width, depth]`, pass `axes=[0, 1, 2]`. |
| 1355 | * for simple batch normalization pass `axes=[0]` (batch only). |
| 1356 | |
| 1357 | Args: |
| 1358 | x: A `Tensor`. |
| 1359 | axes: Array of ints. Axes along which to compute mean and |
| 1360 | variance. |
| 1361 | shift: Not used in the current implementation. |
| 1362 | keepdims: produce moments with the same dimensionality as the input. |
| 1363 | name: Name used to scope the operations that compute the moments. |
| 1364 | |
| 1365 | Returns: |
| 1366 | Two `Tensor` objects: `mean` and `variance`. |
| 1367 | """ |
| 1368 | return moments(x=x, axes=axes, shift=shift, name=name, keep_dims=keepdims) |
| 1369 | |
| 1370 | |
| 1371 | @tf_export(v1=["nn.weighted_moments"]) |