| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
dimension.
The current shape paritition algorithm does not partition the most minor
dimension, because doing so causes dynamic loop bounds for the inner loop and
used to prohibit LLVM vectorization. This constraint has been removed with
revision 328478 and LLVM can now vectorize loops with dynamic bounds. Allow
partitioning the most minor dimension is also necessary to support the
parallelization of matrix-vector multiplication.
Adjust shape_partition_test to reflect this change in the shape partition
algorithm.
PiperOrigin-RevId: 190574615
|
|
|
|
| |
PiperOrigin-RevId: 190573872
|
|
|
|
|
|
| |
of the exposed TF API.
PiperOrigin-RevId: 190570489
|
|
|
|
| |
PiperOrigin-RevId: 190566036
|
|
|
|
| |
PiperOrigin-RevId: 190564824
|
|
|
|
|
|
| |
attributes of dynamic objects, like function calls.
PiperOrigin-RevId: 190564784
|
|
|
|
| |
PiperOrigin-RevId: 190564365
|
|
|
|
| |
PiperOrigin-RevId: 190564036
|
|
|
|
| |
PiperOrigin-RevId: 190563114
|
|
|
|
|
|
| |
of upcoming DistributionStrategy support in Estimator.
PiperOrigin-RevId: 190563074
|
|
|
|
| |
PiperOrigin-RevId: 190561679
|
|
|
|
|
|
| |
New home: https://github.com/tensorflow/probability/tree/master/tensorflow_probability/python/mcmc
PiperOrigin-RevId: 190560180
|
|
|
|
|
|
|
| |
- Add a xla.OpSharding field to the HloInstructionProto.
- Metatdata handling is tested.
PiperOrigin-RevId: 190553731
|
|
|
|
| |
PiperOrigin-RevId: 190548854
|
|
|
|
| |
PiperOrigin-RevId: 190543066
|
|
|
|
|
|
| |
to be consistent with export_saved_model
PiperOrigin-RevId: 190539064
|
|
|
|
| |
PiperOrigin-RevId: 190537320
|
|
|
|
|
|
| |
/ iterator stack.
PiperOrigin-RevId: 190536945
|
|
|
|
| |
PiperOrigin-RevId: 190536468
|
|
|
|
|
|
|
|
| |
Also,
- Templatized ComputeAndCompareRX and CreateRXParameter so that they accept XlaBuilder and XlaOp.
- Clear data held by an XlaBuilder when Build() is called, otherwise errors will occur when the builder is reused.
PiperOrigin-RevId: 190534245
|
|
|
|
| |
PiperOrigin-RevId: 190532168
|
|
|
|
| |
PiperOrigin-RevId: 190526387
|
|
|
|
|
|
| |
`tf.distributions.Distribution`.
PiperOrigin-RevId: 190521666
|
|
|
|
| |
PiperOrigin-RevId: 190517622
|
|
|
|
| |
PiperOrigin-RevId: 190515509
|
|
|
|
| |
PiperOrigin-RevId: 190514839
|
|
|
|
| |
PiperOrigin-RevId: 190512928
|
|
|
|
| |
PiperOrigin-RevId: 190512302
|
|
|
|
| |
PiperOrigin-RevId: 190511964
|
|
|
|
| |
PiperOrigin-RevId: 190508614
|
|
|
|
| |
PiperOrigin-RevId: 190507631
|
|
|
|
| |
PiperOrigin-RevId: 190504933
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fixes: #16619
See the referred GitHub issue for details, but users want to be
able to skip certain nodes when searching for inf/nans, because
some nodes generate inf/nans even in nominal conditions.
This CL adds a new optional flag `--filter_exclude_node_names`
(or `-fenn` for short), which allows users to do exactly that,
by using a regex for node names.
RELNOTES: tfdbg CLI: Allow exclusion of nodes by regular expressions
during tensor filter-enabled Session runs: see the new flags
`--filter_exclude_node_names` (or `-fenn` for short).
PiperOrigin-RevId: 190504225
|
|
|
|
| |
PiperOrigin-RevId: 190503823
|
|
|
|
| |
PiperOrigin-RevId: 190500020
|
|
|
|
|
|
|
|
| |
Requiring it
to happen before OpInit() is way too confusing for users.
PiperOrigin-RevId: 190499644
|
|
|
|
|
|
| |
the presence of ops that modify their inputs in-place: Do not dedup nodes if the underlying buffers for their outputs may be passed to an in-place op.
PiperOrigin-RevId: 190499037
|
|
|
|
| |
PiperOrigin-RevId: 190497916
|
|
|
|
|
|
| |
by tf.GradientTape.gradient() at this time.
PiperOrigin-RevId: 190494436
|
|
|
|
| |
PiperOrigin-RevId: 190493969
|
|
|
|
|
|
|
|
| |
extended BasicDecoder which for example returns a tf.contrib.seq2seq.AttentionWrapperState.
In this case the internal while-loop fails when trying to store an instance tf.contrib.seq2seq.AttentionWrapperState in the internal TensorArray.
PiperOrigin-RevId: 190491787
|
|
|
|
| |
PiperOrigin-RevId: 190489086
|
|
|
|
|
|
| |
reads imagenet TFRecord files.
PiperOrigin-RevId: 190488817
|
|
|
|
|
|
| |
function buffering resource so that we can start using it with re-initializable iterators.
PiperOrigin-RevId: 190484110
|
|
|
|
| |
PiperOrigin-RevId: 190479555
|
|
|
|
|
|
| |
https://www.tensorflow.org/api_docs/python/tf/contrib/training/HParams to show that it allows '=' in the values.
PiperOrigin-RevId: 190470578
|
|
|
|
| |
PiperOrigin-RevId: 190416108
|
|
|
|
| |
PiperOrigin-RevId: 190415923
|
|
|
|
| |
PiperOrigin-RevId: 190411045
|
|
|
|
| |
PiperOrigin-RevId: 190409217
|