- 14 11月, 2019 21 次提交
-
-
由 George Karpenkov 提交于
PiperOrigin-RevId: 280227040 Change-Id: I39854dfbdf998b08c7d276a50cfb5ee4406b861b
-
由 River Riddle 提交于
This refactors the implementation of block signature(type) conversion to not insert fake cast operations to perform the type conversion, but to instead create a new block containing the proper signature. This has the benefit of enabling the use of pre-computed analyses that rely on mapping values. It also leads to a much cleaner implementation overall. The major user facing change is that applySignatureConversion will now replace the entry block of the region, meaning that blocks generally shouldn't be cached over calls to applySignatureConversion. PiperOrigin-RevId: 280226936 Change-Id: Ibb8cb815c59bb1f805377d11bb66c6ff9808ec82
-
由 Suharsh Sivakumar 提交于
Add 'quantization_parameters' field to python interpreter to allow getting the full per-axis quantization parameters via python. PiperOrigin-RevId: 280220657 Change-Id: I3fe27b9c88a717f9c85e59164cd12449679006e3
-
由 Lucy Fox 提交于
This creates a central place in the documentation where MLIR-specific terminology is defined. See discussion on the MLIR forum (https://groups.google.com/a/tensorflow.org/g/mlir/c/5YXDSdu76Hk). PiperOrigin-RevId: 280220365 Change-Id: I370d5bd7bc329b81212d9cca37f107dd0d8243be
-
由 Robert David 提交于
PiperOrigin-RevId: 280220301 Change-Id: Ib36618937995f92d65660d806ecf90e9d608ab9a
-
由 Sergei Lebedev 提交于
PiperOrigin-RevId: 280218429 Change-Id: I51024998ef3f85e1f99fface6f6f36a1443ef031
-
由 River Riddle 提交于
The current implementation silently fails if the '@' identifier isn't present, making it similar to the 'optional' parse methods. This change renames the current implementation to 'Optional' and adds a new 'parseSymbolName' that emits an error. PiperOrigin-RevId: 280214610 Change-Id: I44da6f57992b814f6e503f06c220cd8973d3d013
-
由 A. Unique TensorFlower 提交于
Previously, `get_concrete_function` did not call _garbage_collector.release() so the graph was deleted when the function was deleted. Also, introduce _get_concrete_function_garbage_collected for places that we want to keep the existing behavior. PiperOrigin-RevId: 280213338 Change-Id: Ica498342b61c129132d254adf61e30256b24efe3
-
由 Martin Wicke 提交于
(cf #29871) PiperOrigin-RevId: 280208993 Change-Id: Ia604e0ce832c7fb2adb5c2eff631a7cd1e0c2629
-
由 George Karpenkov 提交于
For some models, resolving constant expressions in the bridge leads to non-linear slowdown. PiperOrigin-RevId: 280205828 Change-Id: Icaa0394d8045c31544b66123857da7cb1b1701b9
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 280204674 Change-Id: Idb551efdde6aa51c2fcfd6fe6e82d175bdcc9bb0
-
由 Jiri Simsa 提交于
PiperOrigin-RevId: 280200721 Change-Id: Icfaffb567b970da140e9b0d3a6c2093452893f01
-
由 Mihai Maruseac 提交于
PiperOrigin-RevId: 280198370 Change-Id: I67e4117575ff3ac00cbda0ac6bfc971c2fbeaf74
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 280185051 Change-Id: I5be656a7a4fa4c6d434b79e59a532dfd8a6dd455
-
由 Benjamin Kramer 提交于
PiperOrigin-RevId: 280184071 Change-Id: I6a0f22809a7d11c5bad7254c20e79525bf3ab88a
-
由 Adrian Kuegel 提交于
It will just be ignored by the starlark rule. PiperOrigin-RevId: 280183097 Change-Id: I456d8a1ef071c5b3a82701cdc4de70d434811cba
-
由 Adrian Kuegel 提交于
Currently both works because of an alias rule, but we want to remove the alias rule. PiperOrigin-RevId: 280163952 Change-Id: Icd6e1938aacf9a76186049558f82f358c7783ea4
-
由 Adrian Kuegel 提交于
xla_proto and hlo_proto are now tf_proto_library_cc targets. PiperOrigin-RevId: 280157517 Change-Id: I51a5c3720856908eb3be4e1b8ea6283d03e7fed6
-
由 Chao Mei 提交于
Move the memory stats output back to BenchmarkModel::Run so that it's always there regardless which benchmark listener is used. PiperOrigin-RevId: 280149185 Change-Id: Ic9932baea72083117e37c75062c2bd9e9c652d9e
-
由 Adrian Kuegel 提交于
PiperOrigin-RevId: 280146999 Change-Id: I6ca6f95e0fc6725bc23823934be1730b19f8478a
-
由 Adrian Kuegel 提交于
This was forgotten when migrating from xla_proto_library to tf_proto_library_cc. PiperOrigin-RevId: 280146847 Change-Id: Ieb837fe06a7af77f559bb287ff22cb9956964876
-
- 13 11月, 2019 19 次提交
-
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 280146467 Change-Id: Ibb580f0f13d9afd4f7c373bbb0eab7022b1a95ce
-
由 Adrian Kuegel 提交于
This was forgotten when migrating from xla_proto_library to tf_proto_library_cc. PiperOrigin-RevId: 280143885 Change-Id: Ibd23bda2239f819dc9ba96bd16e17c42a53ff55a
-
由 Guangda Lai 提交于
failure with TRT 6.0. PiperOrigin-RevId: 280129198 Change-Id: Ife5ee7958c8cdc1438c11e059dd18c834304c9c5
-
由 A. Unique TensorFlower 提交于
Since VariableOp is serialized during processBlock, we add two more fields, `functionHeader` and `functionBody`, to collect instructions for a function. After all the blocks have been processed, we append them to the `functions`. Also, fix a bug in processGlobalVariableOp. The global variables should be encoded into `typesGlobalValues`. PiperOrigin-RevId: 280105366 Change-Id: I5389451312d55869a1f6740d87597444013421b2
-
由 Frank Chen 提交于
PiperOrigin-RevId: 280102761 Change-Id: I057c8074e3e25ee04a44e4ce77fd4981babd17f4
-
由 Penporn Koanantakool 提交于
PiperOrigin-RevId: 280097036 Change-Id: I4578cacb876f3d27bb860c3794b685f2bc26da03
-
由 Smit Hinsu 提交于
PiperOrigin-RevId: 280096904 Change-Id: I0d5e9233da52afd2f7c13094574a59c0d8403403
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 280096487 Change-Id: I506633f77f4042fa19f81c078f307f843a9e0c1f
-
由 Renjie Liu 提交于
PiperOrigin-RevId: 280090706 Change-Id: I628bde953f31500ce3588127c7e42414f002f8b4
-
由 Henry Tan 提交于
PiperOrigin-RevId: 280085436 Change-Id: If02ccdebcddeead5b80f85790a7d7e8a05d30c4d
-
由 Yuefeng Zhou 提交于
PiperOrigin-RevId: 280080043 Change-Id: Id152c198ebf20256fc14b2ea1e16b8c5db71844c
-
由 A. Unique TensorFlower 提交于
For that we need to generate multiple device events (one per device) AND one driver callback events. Add some trace me for nccl_manager.cc PiperOrigin-RevId: 280077382 Change-Id: Ice0ebe4b98f3a6a222e210d266666b0e7fbf8171
-
由 Scott Zhu 提交于
Add 'quantization_parameters' field to python interpreter to allow getting the full per-axis quantization parameters via python. PiperOrigin-RevId: 280077041 Change-Id: I32e7c8053e7550b47af74600a9d5a0d3b01b54bb
-
由 T.J. Alumbaugh 提交于
PiperOrigin-RevId: 280072888 Change-Id: I7293fe0632fea5a29f3c8fa644f1437adc37cdd0
-
由 Rohan Jain 提交于
OwnedIterators and MultiDeviceIterators get created using their components at least twice during function tracing. We don't want to run the deleter in those cases as they are just referring to the original resource and not creating a new one. So we just create the deleter the first time. PiperOrigin-RevId: 280071920 Change-Id: I266ac500246354e6e4d6145835c4f90926ec5715
-
由 Gaurav Jain 提交于
PiperOrigin-RevId: 280070797 Change-Id: I3c8d6ba04e51c05d0ffb515b87b8f686f74c31a0
-
由 Hye Soo Yang 提交于
PiperOrigin-RevId: 280067957 Change-Id: I0ff98b7754b3c4356db1ee32eede9582ab3c7a72
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 280067091 Change-Id: I83939e8da82e9e488685afa36903ec9980dc5b50
-
由 Robert David 提交于
PiperOrigin-RevId: 280065803 Change-Id: I9079f18e1064bf2cef8523d9e3c5f060eeda0a8e
-