提交 · f260cb043a7877657ec0b25d21b023aa3fdcc7cf · qq_38905368 / tensorflow

14 12月, 2018 5 次提交
- A
  Internal change. · f260cb04
  由 A. Unique TensorFlower 提交于 12月 13, 2018
```
PiperOrigin-RevId: 225377288
```
  f260cb04
- S
  Fix instructions for building demo. · 20fee25c
  由 Shashi Shekhar 提交于 12月 13, 2018
```
Add --cxxopt='--std=c++11' to instructions.

PiperOrigin-RevId: 225376880
```
  20fee25c
- D
  Use weak refs when caching symbols in the namespace, to avoid tripping... · de67c554
  由 Dan Moldovan 提交于 12月 13, 2018
```
Use weak refs when caching symbols in the namespace, to avoid tripping circular reference detectors in tests.

PiperOrigin-RevId: 225376019
```
  de67c554
- A
  Add concrete functions to graph. · 04f51da5
  由 A. Unique TensorFlower 提交于 12月 13, 2018
```
PiperOrigin-RevId: 225372823
```
  04f51da5
- A
  [TF:XLA] Use XLA CPU runtime functions to speed up R2 dot in the HLO evaluator. · 4e7413a2
  由 A. Unique TensorFlower 提交于 12月 13, 2018
```
This CL adds a fast-path for R2 dot. For now the fast implementation has certain limitations:

1. Only operands with default layout, and
2. float type

It uses the XLA's CPU runtime functions which invoke eigen.

PiperOrigin-RevId: 225372611
```
  4e7413a2
13 12月, 2018 35 次提交
- S
  Misc. fixes for optional variant tensors. · 3dfbc353
  由 Skye Wanderman-Milne 提交于 12月 13, 2018
```
These are required for cond_v2 to work properly on GPU. The fix in optional_ops.cc is subtle, but I can't think of a way to more generally fix this lifetime issue.

PiperOrigin-RevId: 225369355
```
  3dfbc353
- A
  Update sequence categorical columns to new FeatureColumn API. · 199fe84d
  由 A. Unique TensorFlower 提交于 12月 13, 2018
```
PiperOrigin-RevId: 225366627
```
  199fe84d
- A
  Avoid keyword arguments to core Starlark builtins. · c5002272
  由 A. Unique TensorFlower 提交于 12月 13, 2018
```
PiperOrigin-RevId: 225364007
```
  c5002272
- A
  Directly call python for the nvcc wrapper. · 3169f329
  由 A. Unique TensorFlower 提交于 12月 13, 2018
```
.bat files have an 8k limit on the command line length which we are now hitting.

PiperOrigin-RevId: 225355646
```
  3169f329
- A
  Remove fused_batch_norm_util from default deps of kernels. · 67d2144c
  由 A. Unique TensorFlower 提交于 12月 13, 2018
```
It contains the same source files as fused_batch_norm_op, leading to duplicate
symbol errors in some configurations.

PiperOrigin-RevId: 225350457
```
  67d2144c
- J
  Refactor ListDevices in TF Go API · 045ac6c3
  由 James Keeling 提交于 12月 13, 2018
```
I pull the logic to convert a *C.TF_DeviceList to a []Device into its own function `deviceSliceFromDeviceList`. This is so I can reuse it when wrapping the Eager C API.

PiperOrigin-RevId: 225344064
```
  045ac6c3
- J
  TF Go Wrapper: Add String() method to Device · ca096a32
  由 James Keeling 提交于 12月 13, 2018
```
This is useful when printing devices.

PiperOrigin-RevId: 225343692
```
  ca096a32
- A
  compat: Update forward compatibility horizon to 2018-12-13 · 5dfb096b
  由 A. Unique TensorFlower 提交于 12月 13, 2018
```
PiperOrigin-RevId: 225327770
```
  5dfb096b
- M
  Remove duplicate build flag. · b5969ad5
  由 Michael Reneer 提交于 12月 13, 2018
```
PiperOrigin-RevId: 225323136
```
  b5969ad5
- A
  Automated rollback of commit 221f4d23 · 62bc640c
  由 A. Unique TensorFlower 提交于 12月 13, 2018
```
PiperOrigin-RevId: 225322589
```
  62bc640c
- A
  Update TF 2.0 package to require Estimator 2.0 package. · 6beb8734
  由 Anna R 提交于 12月 13, 2018
```
PiperOrigin-RevId: 225321405
```
  6beb8734
- G
  Add TRT conversion support to saved_model_cli · b99d914c
  由 Guangda Lai 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225318469
```
  b99d914c
- A
  Automated rollback of commit 306cf4b2 · ad26fe70
  由 A. Unique TensorFlower 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225310802
```
  ad26fe70
- A
  Add tpu_ordinal_selector_op. · c2cb414e
  由 A. Unique TensorFlower 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225307980
```
  c2cb414e
- A
  use optimized_ops::TanH directly · 7b9f418c
  由 A. Unique TensorFlower 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225299240
```
  7b9f418c
- Y
  Stop TPU computation if compilation fails. · 7ac4a1e4
  由 Youlong Cheng 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225298488
```
  7ac4a1e4
- A
  Add a config setting to exclude xla from tensorflow_opensource target. · 306cf4b2
  由 A. Unique TensorFlower 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225296419
```
  306cf4b2
- T
  Merge pull request #24078 from trevor-m:tmorris_tftrt_strided_slice_op · 5ce09c99
  由 TensorFlower Gardener 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225296218
```
  5ce09c99
- A
  Go: Update generated wrapper functions for TensorFlow ops. · 86e4902d
  由 A. Unique TensorFlower 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225294738
```
  86e4902d
- S
  Enable recurrent_test to run in v2 mode. · 99afa4b4
  由 Scott Zhu 提交于 12月 12, 2018
```
Also increase the shard to speed up the tests.

PiperOrigin-RevId: 225293834
```
  99afa4b4
- A
  Update ops-related pbtxt files. · 6345287a
  由 A. Unique TensorFlower 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225292379
```
  6345287a
- S
  Enable v2 mode for lstm_test. · 4637671e
  由 Scott Zhu 提交于 12月 12, 2018
```
Test that can only run in v1 has the proper bug attached.

PiperOrigin-RevId: 225291768
```
  4637671e
- A
  Upstream Eigen update to 9f48e814419e823f4a0a31b0cc365aa850c164f5. · 3eb7616b
  由 A. Unique TensorFlower 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225289423
```
  3eb7616b
- F
  Run BatchNorm layer tests in all relevant execution modes. · a3d63443
  由 Francois Chollet 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225287527
```
  a3d63443
- D
  Default `to_code` to brief verbosity (since it returns the code already). · 1ec3b398
  由 Dan Moldovan 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225283401
```
  1ec3b398
- F
  Enable a number of tests for merge layers to run in eager mode. · 00d91e7b
  由 Francois Chollet 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225283392
```
  00d91e7b
- A
  Adds unicode_decode and unicode_decode_with_offsets ops, which decode strings... · c0b2e3eb
  由 A. Unique TensorFlower 提交于 12月 12, 2018
```
Adds unicode_decode and unicode_decode_with_offsets ops, which decode strings into unicode codepoints.

Adds unicode_split and unicode_split_with_offset ops, which split strings into unicode characters.

RELNOTES: Adds unicode_decode, unicode_decode_with_offsets, unicode_split, and unicode_split_with_offset ops.
PiperOrigin-RevId: 225281768
```
  c0b2e3eb
- A
  Fix a bug in lstm_eval. · ca0ccc6f
  由 A. Unique TensorFlower 提交于 12月 12, 2018
```
PiperOrigin-RevId: 225281253
```
  ca0ccc6f
- T
  
  Fix failed narrowing checks · ca10e2d3
  由 Trevor Morris 提交于 12月 12, 2018
  
  ca10e2d3
- T
  Fix bug with masking and undefined batch dims. Masking needs to take place... · 1254fdd1
  由 Trevor Morris 提交于 12月 11, 2018
```
Fix bug with masking and undefined batch dims. Masking needs to take place inside of GetStridedSliceBound
```
  1254fdd1
- T
  
  Apply smit-hinsu's suggestions · 296b83f1
  由 Trevor Morris 提交于 12月 11, 2018
  
  296b83f1
- T
  
  Fix usage of TensorShapeArrayToTrtDims · 661acd19
  由 Trevor Morris 提交于 12月 10, 2018
  
  661acd19
- T
  
  Make unsupported mask options clearer · d16eafc6
  由 Trevor Morris 提交于 12月 05, 2018
  
  d16eafc6
- T
  
  Fix clang-format · 7e17c6af
  由 Trevor Morris 提交于 12月 03, 2018
  
  7e17c6af
- T
  
  Formatting · edccece9
  由 Trevor Morris 提交于 11月 30, 2018
  
  edccece9

qq_38905368 / tensorflow 与 Fork 源项目一致

qq_38905368 / tensorflow
与 Fork 源项目一致