- 17 12月, 2019 40 次提交
-
-
由 Brian Atkinson 提交于
PiperOrigin-RevId: 285896522 Change-Id: I2116497da6a683c5fb62e574794e4b1e04dee571
-
由 Karim Nosir 提交于
PiperOrigin-RevId: 285891574 Change-Id: I67da00e81f6ca60e939b5f77502be1249d59a930
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 285891169 Change-Id: I20b55062c56c5b04c8dc0fc03ad6df6f5f02c8f5
-
由 Karim Nosir 提交于
- Add new method for GenerateTestSpec to generate testspec from TFLite. PiperOrigin-RevId: 285884997 Change-Id: I803bd619013f7410bd56283a715e46c8719d4810
-
由 Mihai Maruseac 提交于
PiperOrigin-RevId: 285882855 Change-Id: I252217310d589d4c331b0287f5fc328d7bb05b75
-
由 Zhenyu Tan 提交于
PiperOrigin-RevId: 285882699 Change-Id: If4d365661a0376ee815fd02fe2a16dd823fca13d
-
由 A. Unique TensorFlower 提交于
third_party/tensorflow/compiler/mlir/tfrt/tfrt_delegate. PiperOrigin-RevId: 285880398 Change-Id: I716a73a66d418a58a6a930175e7673b9b59ae671
-
由 Derek Murray 提交于
PiperOrigin-RevId: 285880109 Change-Id: Ifa5b65fe41bea505d1454a8c43b56959eaa04431
-
由 Yifei Feng 提交于
PiperOrigin-RevId: 285871263 Change-Id: I10b5eacb8861ffc07384ea76e4e4597283f69fcb
-
由 Andy Ly 提交于
This reuses helper functions for calculating begin indices, end indices, and strides from StridedSliceGrad. The unit test for StridedSlice with shrink_axis_mask is removed until helper functions are updated to support shrink_axis_mask. PiperOrigin-RevId: 285870165 Change-Id: I278c023770b767d287bc0de994e7301800638821
-
由 Yunxing Dai 提交于
PiperOrigin-RevId: 285867393 Change-Id: Ia9bbb5933a157c053cd26357160bdb729b7b5f90
-
由 TensorFlower Gardener 提交于
PiperOrigin-RevId: 285864348 Change-Id: I63ba578cbb1a23f776570cc8a4cdb7fe3312a7d4
-
由 Mihai Maruseac 提交于
It used to be transitively brought in by keras_applications but we no longer depend on that. PiperOrigin-RevId: 285864272 Change-Id: Iae2de4f9b8686602d661cff716ff592788b43b09
-
由 Dero Gharibian 提交于
PiperOrigin-RevId: 285863781 Change-Id: Id8c15e610cd15b4ae78d89efe1a5aed9956d6e11
-
由 Smit Hinsu 提交于
ConstantSplat requires static shaped types BinBroadcastDimensions requires ranked types PiperOrigin-RevId: 285863722 Change-Id: Ia2e1220568ab4eae8683b4c7b74ab1c4a38a1240
-
由 Rohan Jain 提交于
PiperOrigin-RevId: 285862836 Change-Id: I2eec29c2300dfbc99f29b30b56e3e7dfea6d047e
-
由 Alexandre Passos 提交于
PiperOrigin-RevId: 285859917 Change-Id: I0176671ce2833cc3d589cbff1b84c9ee89a99361
-
由 Alexandre Passos 提交于
PiperOrigin-RevId: 285859645 Change-Id: Ifd3656ef665ae19591414ab6051dd19a68b99bf4
-
由 Mahesh Ravishankar 提交于
Some changes to the dialect generation script to allow specification of different base class to derive from in ODS. PiperOrigin-RevId: 285859230 Change-Id: I14e96a424e7c63f93435561074ed8b52e4ce78da
-
由 Karim Nosir 提交于
PiperOrigin-RevId: 285858274 Change-Id: I96823afe6d72cfc96aa838f4e2ba3a3d55d2d238
-
由 Rick Chao 提交于
This enables testing of multi-worker fault tolerance where some processes are killed at some point of time while the others continues running. PiperOrigin-RevId: 285850240 Change-Id: I18c99604a9dbb57daf7fb39419ae40db45534523
-
由 Srinivas Vasudevan 提交于
PiperOrigin-RevId: 285849818 Change-Id: Ida006281680c9eccfc68b24a455cfe578183c6e4
-
由 Mahesh Ravishankar 提交于
Scope and Memory Semantics attributes need to be serialized as a constant integer value and the <id> needs to be used to specify the value. Fix the auto-generated SPIR-V (de)serialization to handle this. PiperOrigin-RevId: 285849431 Change-Id: I6946cf3622fd52b4231ea944cfef5528adba25af
-
由 TensorFlower Gardener 提交于
PiperOrigin-RevId: 285846177 Change-Id: I95e3390170aa2f6badc183ca44bd5f580894ef85
-
由 Peng Wang 提交于
Adds examples in stateful_random_ops.py, improves some docstrings, and also adds missing export for algorithm enum. PiperOrigin-RevId: 285845389 Change-Id: Iaa60466151ba761a6f5a1ef6cd8b34cbbfdde5b1
-
由 Zhenyu Tan 提交于
PiperOrigin-RevId: 285843812 Change-Id: Ifd8183f916d673a7e661ca5dca84312c5cc74078
-
由 A. Unique TensorFlower 提交于
targets and tests from tensorflow/core to tensorflow/core/util. PiperOrigin-RevId: 285843378 Change-Id: Ibeece4c1d3ab69cb6805790f369d6fc04ac434e3
-
由 Smit Hinsu 提交于
PiperOrigin-RevId: 285841356 Change-Id: I37d1eb259bc875cdfd8a53ad8df63706e761fbb2
-
由 Vishnuvardhan Janapati 提交于
-
由 Nicolas Vasilache 提交于
This CL adds more Linalg EDSC ops and tests to support building pointwise operations along with conv and dilated_conv. This also fixes a bug in the existing linalg_matmul EDSC and beefs up the test. The current set of ops is already enough to build an interesting, albeit simple, model used internally. PiperOrigin-RevId: 285838012 Change-Id: I35edf4bed5eef32a22900c89c4482f5426b4645d
-
由 Smit Hinsu 提交于
PiperOrigin-RevId: 285837508 Change-Id: Ie870171df441467fd3cc06cfd5a2dcacd341437c
-
由 TensorFlower Gardener 提交于
PiperOrigin-RevId: 285834307 Change-Id: Ib8140e45fd45ee30ddc365951fe843469cbcb471
-
由 Smit Hinsu 提交于
PiperOrigin-RevId: 285832508 Change-Id: I35f79b5a6d8390bd62c4ae4e53fa5c5c4da30d2d
-
由 Smit Hinsu 提交于
PiperOrigin-RevId: 285831517 Change-Id: I40fea9dd6a2b3b1f0d71249b5f28221acd648704
-
由 Srinivas Vasudevan 提交于
- Removed restriction on tridiagonal_matmul and tridiagonal_solve for shapes to be statically known. PiperOrigin-RevId: 285830396 Change-Id: I2291ee1111f2a41e658b085cbd402d2977acb939
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 285830394 Change-Id: Iea5547d55efedd5150c82edf7afb5e4ed2b373d3
-
由 Smit Hinsu 提交于
Also, add custom verifier for the DynamicStitch op. PiperOrigin-RevId: 285828224 Change-Id: I8983809c5d011672673e167b9fa2dc1bef0bf413
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 285822581 Change-Id: I679256cc6f5890fa93ff3a2bfb9136b5d679d3ac
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 285822505 Change-Id: Ieb1d4af1df36016e6ecb919c32334c84d2327453
-
由 A. Unique TensorFlower 提交于
This updates the lowering pipelines from the GPU dialect to lower-level dialects (NVVM, SPIRV) to use the recently introduced gpu.func operation instead of a standard function annotated with an attribute. In particular, the kernel outlining is updated to produce gpu.func instead of std.func and the individual conversions are updated to consume gpu.funcs and disallow standard funcs after legalization, if necessary. The attribute "gpu.kernel" is preserved in the generic syntax, but can also be used with the custom syntax on gpu.funcs. The special kind of function for GPU allows one to use additional features such as memory attribution. PiperOrigin-RevId: 285822272 Change-Id: I35cbe2454e0f43baa4e2c4c16434ae2158f32d38
-