- 13 8月, 2019 15 次提交
-
-
由 Dero Gharibian 提交于
This is a part of a larger migration effort for tensorflow::tstring. See: https://github.com/tensorflow/community/pull/91 PiperOrigin-RevId: 262957475
-
由 Tiezhen WANG 提交于
PiperOrigin-RevId: 262956765
-
由 Feng Liu 提交于
This build target is very useful for prototyping tools with GraphDebugInfo. PiperOrigin-RevId: 262955134
-
由 Yunxing Dai 提交于
PiperOrigin-RevId: 262951106
-
由 Gaurav Jain 提交于
Fixes #9359 PiperOrigin-RevId: 262948811
-
由 Priya Gupta 提交于
PiperOrigin-RevId: 262947974
-
由 Pete Warden 提交于
PiperOrigin-RevId: 262947132
-
由 Tiezhen WANG 提交于
This is to mimic building environment on CHRE. PiperOrigin-RevId: 262945402
-
由 Feng Liu 提交于
PiperOrigin-RevId: 262943818
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 262942762
-
由 Zhuo Peng 提交于
PiperOrigin-RevId: 262940594
-
由 TensorFlower Gardener 提交于
PiperOrigin-RevId: 262939740
-
由 Jacques Pienaar 提交于
Prefer to enumerate all cases in the switch instead of using default to allow compiler to flag missing cases. This also avoids -Wcovered-switch-default warning. PiperOrigin-RevId: 262935972
-
由 Gaurav Jain 提交于
When comparing CacheKeys we cannot simply compare the namedtuple fields since that might involve comparing 2 Variables whose equality comparison returns a Tensor rather than a bool. We thus compare Variables using their class name, dtype & shape, similar to the processing we do for the hash. PiperOrigin-RevId: 262935453
-
由 Jacques Pienaar 提交于
This can result in index expression overflow in "Loc.getPointer() - ColumnNo" in SourgeMgr. loc could also be prefixed to the message additionally in this case. PiperOrigin-RevId: 262935408
-
- 12 8月, 2019 25 次提交
-
-
由 Alex Stark 提交于
PiperOrigin-RevId: 262926213
-
由 Lasse Espeholt 提交于
PiperOrigin-RevId: 262925866
-
由 Dero Gharibian 提交于
This is a part of a larger migration effort for tensorflow::tstring. See: https://github.com/tensorflow/community/pull/91 PiperOrigin-RevId: 262918920
-
由 Dero Gharibian 提交于
This is a part of a larger migration effort for tensorflow::tstring. See: https://github.com/tensorflow/community/pull/91 PiperOrigin-RevId: 262918772
-
由 A. Unique TensorFlower 提交于
Every new instance of TfLiteGlCompileOptions should be populated through DefaultTfLiteGlCompileOptions() call to avoid breaking changes when new option is introduced. PiperOrigin-RevId: 262918305
-
由 A. Unique TensorFlower 提交于
This instruction is a local counterpart of llvm.global that takes a symbol reference to a global and produces an SSA value containing the pointer to it. Used in combination, these two operations allow one to use globals with other operations expecting SSA values. At a cost of IR indirection, we make sure the functions don't implicitly capture the surrounding SSA values and remain suitable for parallel processing. PiperOrigin-RevId: 262908622
-
由 Nicolas Vasilache 提交于
This CL is step 3/n towards building a simple, programmable and portable vector abstraction in MLIR that can go all the way down to generating assembly vector code via LLVM's opt and llc tools. This CL adds support for converting MLIR n-D vector types to (n-1)-D arrays of 1-D LLVM vectors and a conversion VectorToLLVM that lowers the `vector.extractelement` and `vector.outerproduct` instructions to the proper mix of `llvm.vectorshuffle`, `llvm.extractelement` and `llvm.mulf`. This has been independently verified to produce proper avx2 code. Input: ``` func @vec_1d(%arg0: vector<4xf32>, %arg1: vector<8xf32>) -> vector<8xf32> { %2 = vector.outerproduct %arg0, %arg1 : vector<4xf32>, vector<8xf32> %3 = vector.extractelement %2[0 : i32]: vector<4x8xf32> return %3 : vector<8xf32> } ``` Command: ``` mlir-opt vector-to-llvm.mlir -vector-lower-to-llvm-dialect --disable-pass-threading | mlir-opt -lower-to-cfg -lower-to-llvm | mlir-translate --mlir-to-llvmir | opt -O3 | llc -O3 -march=x86-64 -mcpu=haswell -mattr=fma,avx2 ``` Output: ``` vec_1d: # @vec_1d # %bb.0: vbroadcastss %xmm0, %ymm0 vmulps %ymm1, %ymm0, %ymm0 retq ``` PiperOrigin-RevId: 262895929
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 262893594
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 262886695
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 262883920
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 262881732
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 262881731
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 262876195
-
由 Chao Mei 提交于
PiperOrigin-RevId: 262872798
-
由 TensorFlower Gardener 提交于
PiperOrigin-RevId: 262870851
-
由 Chao Mei 提交于
PiperOrigin-RevId: 262869158
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 262866402
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 262864322
-
由 Taehee Jeong 提交于
PiperOrigin-RevId: 262861510
-
由 A. Unique TensorFlower 提交于
failures. PiperOrigin-RevId: 262857756
-
由 Feng Liu 提交于
PiperOrigin-RevId: 262855058
-
由 Dero Gharibian 提交于
This is a part of a larger migration effort for tensorflow::tstring. See: https://github.com/tensorflow/community/pull/91 PiperOrigin-RevId: 262850088
-
由 River Riddle 提交于
The pattern list is not modified by any of these APIs and should thus be passed with const. PiperOrigin-RevId: 262844002
-
由 A. Unique TensorFlower 提交于
1/8/16 bit attrs. NFC PiperOrigin-RevId: 262843016
-
由 Yanan Cao 提交于
Add tf_device dialect that models TensorFlow's actions of launching computations on accelerator devices. PiperOrigin-RevId: 262833506
-