| Commit message (Collapse) | Author | Age |
|
|
|
|
|
| |
data_format attr.
PiperOrigin-RevId: 214608039
|
|
|
|
| |
PiperOrigin-RevId: 214587760
|
|
|
|
| |
PiperOrigin-RevId: 214575129
|
|
|
|
| |
PiperOrigin-RevId: 214574383
|
|
|
|
| |
PiperOrigin-RevId: 214569615
|
|
|
|
| |
PiperOrigin-RevId: 214557082
|
|
|
|
| |
PiperOrigin-RevId: 214553359
|
|
|
|
|
|
| |
are assumed to be the same for HandleImag and HandleReal, when in fact they should be float and complex64 (or float for HandleReal's case), respectively.
PiperOrigin-RevId: 214548051
|
|
|
|
|
|
|
|
| |
supported in Graph mode using initializable iterators. In a subsequent change, we'll add in support for Eager mode as well.
This removes prefetching_ops_v2 code.
PiperOrigin-RevId: 214546754
|
|
|
|
|
|
|
| |
These ops were never intended to support integer dtypes, and the
OpKernels have already been removed in a previous patch.
PiperOrigin-RevId: 214542750
|
|
|
|
| |
PiperOrigin-RevId: 214542049
|
|
|
|
| |
PiperOrigin-RevId: 214541221
|
|
|
|
| |
PiperOrigin-RevId: 214532827
|
|
|
|
|
|
|
|
| |
The original CL broke the InterpreterTest due to use of a newly
deprecated API. This has been fixed, and deprecated API usage in the
samples has also been updated.
PiperOrigin-RevId: 214532691
|
|
|
|
|
|
|
|
| |
Bazel does not allow python rules to directly depend on c++ rules.
So I have to separately manage static dependencies, unfortunately avoiding
"kernels" option for now.
PiperOrigin-RevId: 214532631
|
|
|
|
| |
PiperOrigin-RevId: 214532043
|
|
|
|
| |
PiperOrigin-RevId: 214528768
|
|
|
|
| |
PiperOrigin-RevId: 214521486
|
|
|
|
|
|
| |
match convolution forward + relu.
PiperOrigin-RevId: 214521083
|
|
|
|
|
|
| |
Synchronize tf-nightly dep on current tb-nightly: https://pypi.org/project/tb-nightly/1.12.0a20180925/
PiperOrigin-RevId: 214521033
|
|
|
|
| |
PiperOrigin-RevId: 214519671
|
|
|
|
| |
PiperOrigin-RevId: 214515610
|
|
|
|
|
|
|
|
| |
distribution from tf/contrib to tfp.
Relevant to tensorflow/tensorflow#21909
PiperOrigin-RevId: 214511101
|
|
|
|
| |
PiperOrigin-RevId: 214507546
|
|
|
|
| |
PiperOrigin-RevId: 214505566
|
|\
| |
| |
| | |
PiperOrigin-RevId: 214504400
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214499338
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214499034
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214495925
|
| |
| |
| |
| |
| |
| | |
output and targets and produces the same loss and metrics.
PiperOrigin-RevId: 214494877
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214493515
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214489904
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214489542
|
| |
| |
| |
| |
| |
| | |
SnapshotResourceVariables function.
PiperOrigin-RevId: 214488033
|
| |
| |
| |
| |
| |
| |
| |
| | |
- Update the flatbuffer download URL to use a pinned version.
- Had to provide a mirror url that doesn't exist, since it is required
by the validation flow, the flatbuffer version will be added later.
PiperOrigin-RevId: 214487576
|
| |
| |
| |
| |
| |
| | |
- Move the benchmarks doc to a different file.
PiperOrigin-RevId: 214483393
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
length" is defined:
* BYTE: The number of bytes in each string. (Default)
* UTF8: The number of UTF-8 encoded Unicode code points in each string.
RELNOTES: Add option to calculate string length in Unicode characters
PiperOrigin-RevId: 214478470
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214478085
|
| |
| |
| |
| |
| |
| | |
failing when running with P100.
PiperOrigin-RevId: 214477405
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214476713
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214476625
|
| |
| |
| |
| |
| |
| | |
Add translation.yaml config for site
PiperOrigin-RevId: 214473776
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214463446
|
|\ \
| | |
| | |
| | | |
PiperOrigin-RevId: 214462540
|
| | |
| | |
| | |
| | |
| | |
| | | |
I don't think this annoyed anyone else yet, it's just a nit I noticed while making sure variables can be garbage collected when tracked via tf.keras.
PiperOrigin-RevId: 214462105
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 214461578
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* Use a FlatMap for instruction_iterators_, and actually remove elements from it (which is cheap for a FlatMap).
* Use the size of the map (which is O(1)) rather than the size of the list (which is O(n)) for instruction_count().
PiperOrigin-RevId: 214459259
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 214451901
|
| | |
| | |
| | |
| | |
| | |
| | | |
AsControlDependency accepts a node name not a tensor name.
PiperOrigin-RevId: 214451885
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The conversion process fails for graphs that use tf.boolean_mask(..., axis=0) --
this op calls tf.concat with an empty array as the first argument.
PiperOrigin-RevId: 214451470
|