| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
| |
- Don't need a test_session() when using
@test_util.run_in_graph_and_eager_modes
- Can use @test_util.run_all_in_graph_and_eager_modes when all tests in the
TestCase should be run in both modes.
PiperOrigin-RevId: 209968609
|
|
|
|
| |
PiperOrigin-RevId: 209968387
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
consistent.
Consider only real uses of fusion parameters when computing buffer interference
using HloDataflowAnalysis.
This is a step in unifying our HLO dataflow framework
to a single source of truth and improves HloOrdering::LiveRangeStrictlyBefore
to match the behavior of BufferLiveness::live_range_strictly_before (which uses
TuplePointsToAnalysis)
PiperOrigin-RevId: 209966644
|
|
|
|
| |
PiperOrigin-RevId: 209966634
|
|
|
|
| |
PiperOrigin-RevId: 209962361
|
|
|
|
|
|
| |
variable.
PiperOrigin-RevId: 209962144
|
|
|
|
| |
PiperOrigin-RevId: 209960839
|
|
|
|
|
|
|
| |
Unfortunately this has to be one big patch, because e.g. absl::StrCat
doesn't accept a TF StringPiece, but as soon as we switch to
absl::string_view, we have to switch away from all of the TF functions.
PiperOrigin-RevId: 209957896
|
|
|
|
|
|
| |
Otherwise executing the op behaves differently from constant folding it.
PiperOrigin-RevId: 209949852
|
|
|
|
| |
PiperOrigin-RevId: 209945757
|
|
|
|
|
|
|
|
|
| |
In the process removes unnecessary restriction on it being called multiple
times on the same variables; it might be necessary to do so if you want to
call ema.apply with different control dependencies in different calls to
session.run, for example.
PiperOrigin-RevId: 209945355
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously the domain instructions inserted before and after an `n`
element tuple required `O(n^2)` memory (and compute) because every
operand and user had its own domain instruction with a tuple sharding
and tuple shape for the exit domains what constructed `n` HloSharding
and `n` Shape proto per domain.
After this change we keep track of the domain instructions inserted and
if we already have a domain instruction with the correct operand and
metadata then we re-use it instead of creating a new one.
Additionally we change HloInstruction and ShardingMetadata to store a
std::shared_ptr to HloSharding so the same instance can be shared by
many instructions. This CL doesn't update all uses to remove all of the
duplicated HloShardings but handles the most wastful cases to reduce
memory usage.
PiperOrigin-RevId: 209924260
|
|
|
|
| |
PiperOrigin-RevId: 209908710
|
|
|
|
|
|
| |
whenever one of the template arguments is not a POD, in particular for types such as string. A memset on a struct that contains a string invalidates the internals of that class (in particular the pointer that stores the characters of the string) and causes segfaults. This allows using weighted quantiles beyond int/float like template parameters. The upstream issue with memcopy of alignment bytes used for padding has been solved separately in the meantime.
PiperOrigin-RevId: 209906298
|
|
|
|
|
|
| |
DistributionStrategy.
PiperOrigin-RevId: 209891724
|
|
|
|
| |
PiperOrigin-RevId: 209884688
|
|
|
|
|
|
| |
This allows MirroredStrategy with one device and between-graph strategies with one tower on each worker.
PiperOrigin-RevId: 209884569
|
|
|
|
|
|
| |
coordinator in CollectiveAllReduceStrategy.
PiperOrigin-RevId: 209879722
|
|
|
|
| |
PiperOrigin-RevId: 209875487
|
|
|
|
| |
PiperOrigin-RevId: 209873671
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Adds an accuracy tool that can be used to develop evaluation pipelines to
evaluate model accuracies.
- The binary can be compiled for mobile platforms and
the tool can be used to evaluate accuracy of models by running the binary on
device.
- Adds an example implementation for imagenet ILSVRC classification task
accuracy evaluation.
- More documentations and details coming soon.
PiperOrigin-RevId: 209869774
|
|
|
|
|
|
|
|
| |
instructions.
This will make adding collective instructions slightly easier.
PiperOrigin-RevId: 209864869
|
|
|
|
| |
PiperOrigin-RevId: 209864694
|
|
|
|
| |
PiperOrigin-RevId: 209864144
|
|
|
|
| |
PiperOrigin-RevId: 209860270
|
|
|
|
| |
PiperOrigin-RevId: 209853753
|
|
|
|
|
|
| |
deterministic.
PiperOrigin-RevId: 209852547
|
|\
| |
| |
| | |
PiperOrigin-RevId: 209850481
|
| |
| |
| |
| |
| |
| | |
Apparently it's not inherited from its py_binary dep for pip tests
PiperOrigin-RevId: 209848894
|
| |
| |
| |
| |
| |
| | |
coordinator in MirroredStrategy.
PiperOrigin-RevId: 209848375
|
| |
| |
| |
| | |
PiperOrigin-RevId: 209847371
|
| |
| |
| |
| | |
PiperOrigin-RevId: 209847352
|
| |
| |
| |
| | |
PiperOrigin-RevId: 209845613
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
input datasets are of type dictionaries with different set of key(s).
FIXES #20626
REL_NOTES: bug fix in `tf.data.Dataset.concatenate()`, now throws error if trying to concatenate two datasets of type dictionaries with different set of key(s).
PiperOrigin-RevId: 209845337
|
| |
| |
| |
| | |
PiperOrigin-RevId: 209845109
|
| |
| |
| |
| | |
PiperOrigin-RevId: 209842403
|
| |
| |
| |
| | |
PiperOrigin-RevId: 209840812
|
| |
| |
| |
| | |
PiperOrigin-RevId: 209839159
|
| |
| |
| |
| |
| |
| |
| |
| | |
self.test_session() has been deprecated in 9962eb5e84b15e309410071b06c2ed2d6148ed44 as its name confuses readers of the test. Moving to cached_session() instead which is more explicit about:
* the fact that the session may be reused.
* the session is not closed even when doing a "with self.test_session()" statement.
PiperOrigin-RevId: 209839057
|
| |
| |
| |
| |
| |
| |
| |
| | |
self.test_session() has been deprecated in 9962eb5e84b15e309410071b06c2ed2d6148ed44 as its name confuses readers of the test. Moving to cached_session() instead which is more explicit about:
* the fact that the session may be reused.
* the session is not closed even when doing a "with self.test_session()" statement.
PiperOrigin-RevId: 209839048
|
| |
| |
| |
| |
| |
| |
| |
| | |
self.test_session() has been deprecated in 9962eb5e84b15e309410071b06c2ed2d6148ed44 as its name confuses readers of the test. Moving to cached_session() instead which is more explicit about:
* the fact that the session may be reused.
* the session is not closed even when doing a "with self.test_session()" statement.
PiperOrigin-RevId: 209839032
|
| |
| |
| |
| |
| |
| |
| |
| | |
self.test_session() has been deprecated in 9962eb5e84b15e309410071b06c2ed2d6148ed44 as its name confuses readers of the test. Moving to cached_session() instead which is more explicit about:
* the fact that the session may be reused.
* the session is not closed even when doing a "with self.test_session()" statement.
PiperOrigin-RevId: 209839030
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
In the recent change to instantiate functions at iterator
initialization time, I forgot to make the corresponding change was to
MultiDeviceIterator. This change fixes that problem and unbreaks
MultiDeviceIterator.
PiperOrigin-RevId: 209838406
|
| |
| |
| |
| |
| |
| |
| |
| | |
context.executing_eagerly() in optimizer.variables()
Remove global collection usage from the Keras model to estimator flow.
PiperOrigin-RevId: 209837803
|
| |
| |
| |
| |
| |
| |
| |
| | |
self.test_session() has been deprecated in 9962eb5e84b15e309410071b06c2ed2d6148ed44 as its name confuses readers of the test. Moving to cached_session() instead which is more explicit about:
* the fact that the session may be reused.
* the session is not closed even when doing a "with self.test_session()" statement.
PiperOrigin-RevId: 209837298
|
| |
| |
| |
| |
| |
| |
| |
| | |
self.test_session() has been deprecated in 9962eb5e84b15e309410071b06c2ed2d6148ed44 as its name confuses readers of the test. Moving to cached_session() instead which is more explicit about:
* the fact that the session may be reused.
* the session is not closed even when doing a "with self.test_session()" statement.
PiperOrigin-RevId: 209836477
|
| |
| |
| |
| |
| |
| | |
which will replace dataset.map(parsing_ops.parse_example(..)).
PiperOrigin-RevId: 209836033
|
| |
| |
| |
| | |
PiperOrigin-RevId: 209835865
|
| |
| |
| |
| | |
PiperOrigin-RevId: 209833197
|
| |
| |
| |
| |
| |
| | |
ops when we have nodes with no examples.
PiperOrigin-RevId: 209830550
|