| Commit message (Collapse) | Author | Age |
... | |
| |
| |
| |
| | |
PiperOrigin-RevId: 214967868
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214964988
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214964640
|
| |
| |
| |
| |
| |
| | |
with associated stats_aggregator.
PiperOrigin-RevId: 214961678
|
| |
| |
| |
| |
| |
| | |
We will re-enable it when it is more robust.
PiperOrigin-RevId: 214956066
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214952666
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
dependency
only to the sparse operators that need it.
Automated rollback of commit 120620caf23a044b8aa2db6ba5984384ec936009
PiperOrigin-RevId: 214950946
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214949709
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Fixes an issue where losses created while executing eagerly were returned as unevaluated lambdas in a defun.
Lazily evaluates Layer losses by default when possible. Even when graph building this is generally a better thing to do (e.g. losses called in a while_loop).
Allows calls to Layer.add_loss when executing eagerly, but only for losses which are not conditional on inputs (no activity regularizers).
PiperOrigin-RevId: 214947108
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214946257
|
| |
| |
| |
| |
| |
| | |
function.
PiperOrigin-RevId: 214945748
|
| |
| |
| |
| |
| |
| |
| | |
Building binaries with XLA support does not enable it by default, it
simply makes it accessible via default binary builds.
PiperOrigin-RevId: 214942824
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214941829
|
| |
| |
| |
| |
| |
| |
| |
| | |
NOTE: All ops and kernels previously previously defined in
tensorflow/contrib/data have had their name prefixed with
"Experimental" to indicate that they are not (yet) stable, and thus
not subject to backwards or forwards compatibility guarantees.
PiperOrigin-RevId: 214940819
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214940748
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
with the case when the background thread terminated because the iterator
finished and yet some other requests were coming in.
1. The GetNextFromShard would see an empty buffer and return cancelled instead
of OutOfRange errors
2. On shutdown, we weren't calling all the pending callbacks.
Tested with runs_per_test=5000
PiperOrigin-RevId: 214939274
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214932861
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214904795
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214895147
|
| |
| |
| |
| |
| |
| | |
Keras and DistributionStrategy
PiperOrigin-RevId: 214890580
|
| |
| |
| |
| |
| |
| | |
for supported operating systems.
PiperOrigin-RevId: 214886845
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214886657
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214883775
|
| |
| |
| |
| |
| |
| |
| |
| | |
distribution strategies. That is always the appropriate option.
In the existing code, we would set it to a partially specified "worker" name that was ambiguous and end up on the GPU.
PiperOrigin-RevId: 214882658
|
| |
| |
| |
| |
| |
| |
| | |
In subsequent change I'll change canned estimators to support FeatureColumn V2
and use this LinearModel.
PiperOrigin-RevId: 214882241
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214879388
|
| |
| |
| |
| |
| |
| | |
tpu_embedding_configuration_py, and tpu_embedding_output_layout_py.
PiperOrigin-RevId: 214879168
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214878428
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214878220
|
|\ \
| | |
| | |
| | | |
PiperOrigin-RevId: 214877195
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 214875840
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
reshape can't determine the size of the last dimension when reshaping
shape (0) to (0, 1).
PiperOrigin-RevId: 214872677
|
|\ \ \
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 214869987
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 214867453
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 214866490
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 214863042
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 214862838
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Some APIs are moving out of core TF repo. These APIs will have their own
API compat tests. Adding flag --only_test_core_api=true which will not
check for changes to non-core APIs.
PiperOrigin-RevId: 214860984
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
xla.estimator_model_fn
PiperOrigin-RevId: 214860981
|
|\ \ \ \
| | | | |
| | | | |
| | | | | |
PiperOrigin-RevId: 214860966
|
| | | | |
| | | | |
| | | | |
| | | | | |
PiperOrigin-RevId: 214854303
|
| | | | |
| | | | |
| | | | |
| | | | | |
PiperOrigin-RevId: 214853860
|
| | | | |
| | | | |
| | | | |
| | | | | |
PiperOrigin-RevId: 214853846
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
the most costly functions in Grappler.
PiperOrigin-RevId: 214853009
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
(finite) dataset to a single element.
PiperOrigin-RevId: 214852364
|
| | | | |
| | | | |
| | | | |
| | | | | |
PiperOrigin-RevId: 214849978
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Revert #20539.
PiperOrigin-RevId: 214849875
|
| | | | |
| | | | |
| | | | |
| | | | | |
PiperOrigin-RevId: 214848216
|
| | | | |
| | | | |
| | | | |
| | | | | |
PiperOrigin-RevId: 214848057
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
See
https://github.com/apple/swift/pull/19588/files#diff-923cd5ac82727b31d446c23641b3d749
for an example usage.
Also removed an experimental API that's no longer needed.
PiperOrigin-RevId: 214847132
|