- 23 6月, 2020 40 次提交
-
-
由 Meghna Natraj 提交于
PiperOrigin-RevId: 317762428 Change-Id: I075791cad44c6d9d4e747f76b0cb50a5d422fe87
-
由 Berkin Ilbeyi 提交于
PiperOrigin-RevId: 317761146 Change-Id: Ia27b6765a547451575240e76ea419f705e29bd32
-
由 Xinyi Wang 提交于
PiperOrigin-RevId: 317760674 Change-Id: Ib7e0adbf4f8f013f21faef07ed4961c078806093
-
由 TensorFlower Gardener 提交于
PiperOrigin-RevId: 317757557 Change-Id: I0a0f0cc9025db7d1fbc7975b07d7d934c6fa8c2f
-
由 Feng Liu 提交于
After this step, all the tfr ops should be raised to the tf ops PiperOrigin-RevId: 317756602 Change-Id: I15731f231caf6c47eac1719fc1c2ed3d01cff515
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 317756434 Change-Id: I14911d5f460131f4b9bdf6031cc7011ddd6a1d53
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 317755766 Change-Id: I105b220b0760792d9f4fa7c1390c1477cdb1c627
-
由 TensorFlower Gardener 提交于
PiperOrigin-RevId: 317755695 Change-Id: Ie58da8ffd1605ee7ce31d68bbd103dd0daf71e78
-
由 Peng Wang 提交于
Adds NumPy LICENSE and lists of NumPy API symbols that `tf.experimental.numpy` implements to tensorflow/third_party/py/numpy/ PiperOrigin-RevId: 317755065 Change-Id: I0efd21b4bf917b7f14ec30e0b5710954844de448
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 317754511 Change-Id: I6bcabccaa626039c53269bf7e8837fef972bd537
-
由 Ahmed S. Taei 提交于
PiperOrigin-RevId: 317753780 Change-Id: I33c44d6355cca3b9a054773afad47b6d00ba6d9f
-
由 Dan Moldovan 提交于
Add a workaround for slowness of tf.while_loop in the default executor when maximum_iterations is set. Fixes #40517. PiperOrigin-RevId: 317753132 Change-Id: I7b5767019ceebd3f21975a990e7c2c05dd878ca6
-
由 Raman Sarokin 提交于
Texture2D implements GPUObject. PiperOrigin-RevId: 317752276 Change-Id: Id1c66b2f6ca7b70475cc82abc422935fc3f1a251
-
由 Yujing Zhang 提交于
If a packed variable is passed to a DatasetOp through a captured function, 1) function instantiation (graph expansion) would insert a PackOp between per-replica arg nodes and the DatasetOp. 2) function execution would feed unpacked ResourceHandles to corresponding sub-functions. PiperOrigin-RevId: 317752101 Change-Id: I53f7f1d6d075fd40a76a32f329f2dcbbf1db6494
-
由 Raman Sarokin 提交于
PiperOrigin-RevId: 317752091 Change-Id: I4e23ab1a9a943c197db7a7be25e5272114fd43c3
-
由 Raman Sarokin 提交于
Added support of BHWDC/HWDC layouts. PiperOrigin-RevId: 317751701 Change-Id: I0cb91a6a2317e13624189081eb321a1562810b15
-
由 Tim Shen 提交于
LHLO uses output-parameter, but TupleSelectOp outputs into a tuple on the device. The current type constraints are wrong, and there is not enough expressiveness in LHLO to define a device-memory representation that can be passed in between kernel launches. PiperOrigin-RevId: 317749542 Change-Id: I6d3350cbf9decf006f239a7208f6bbef0175ac61
-
由 Tiezhen WANG 提交于
PiperOrigin-RevId: 317748628 Change-Id: I6822e5991d892d238acb0976966e90f60e130961
-
由 Andy Ly 提交于
tf.BesselI0e and tf.BesselI1e are moved to tf_ops.td as they are now python generated ops, so they do not have summaries and descriptions in the op registry. int8 and int16 support has been added to tf.Acos, tf.Atan, tf.Inv, tf.Reciprocal, tf.Round, and tf.Tan. PiperOrigin-RevId: 317748438 Change-Id: Icb86560d5118c38c69819fd02daa3a3841e113b2
-
由 A. Unique TensorFlower 提交于
Adds some interoperation testing for ndarray methods. Improves support for tf.Variable. Points np.__doc__ to README.md PiperOrigin-RevId: 317747715 Change-Id: Ia19bc25959dd5ddb00f828c5aaa17ac5a35c8131
-
由 Rick Chao 提交于
PiperOrigin-RevId: 317747609 Change-Id: I8bf2e493431a69a0cf581012666045df2879055e
-
由 Raman Sarokin 提交于
PiperOrigin-RevId: 317746681 Change-Id: I8471f352e2f842d40d71c2733564ab346d4eb69c
-
由 Raman Sarokin 提交于
PiperOrigin-RevId: 317746588 Change-Id: I9bd2eb57a3a39db01636f0e12ab812d5c21bf251
-
由 Raman Sarokin 提交于
Merged 2D/3D versions. PiperOrigin-RevId: 317746537 Change-Id: I69961e9ae1412ccff177a44493020ef97ee0921f
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 317746384 Change-Id: I437965917d7d392b906cecd145f87ced4635a592
-
由 Raman Sarokin 提交于
Buffer implements GPUObject. PiperOrigin-RevId: 317746356 Change-Id: I715d37a924298e54e9fe68e5c7b719ab592237a1
-
由 Zhenyu Tan 提交于
PiperOrigin-RevId: 317744390 Change-Id: I62557f99761043c3e3c83de4f4faee0cb66d94a8
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 317743030 Change-Id: I8ac4e0c30cb67cfc611d351ba405e2b94e57c04a
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 317742833 Change-Id: Ib78b44b673b29dd3ff0b81fa00068f64198feb30
-
由 Akshay Modi 提交于
functions PiperOrigin-RevId: 317741132 Change-Id: I0c1ba744b0df7a648879da58bd600e81e26184ee
-
由 Yanhua Sun 提交于
DeviceIndex op: given a list of device names, this operation returns the index of the device this op runs. In the case of XLA, we are not executing on any device, we return the length of the list. PiperOrigin-RevId: 317740778 Change-Id: I0679aa0adc5508b83502eee0d2044584577ed5b4
-
由 TensorFlower Gardener 提交于
PiperOrigin-RevId: 317740303 Change-Id: Ice0d235e04932eb44f81bd5029cbdbf6279abe2c
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 317738384 Change-Id: I2c74c34bb5fed2809a46f4ecdb95616f2f1c832d
-
由 Raman Sarokin 提交于
PiperOrigin-RevId: 317737917 Change-Id: Ie21457323a2cdb0376c16065fc4b809a5cc9a1d3
-
由 Rick Chao 提交于
PiperOrigin-RevId: 317736970 Change-Id: Ia9c76462afc4c2fcc42a149960b50b2cbcafd482
-
由 Smit Hinsu 提交于
- Add PreluTester class and unit test for XNNPACK-delegated Prelu operator - Relax restrictions on the number of input/output dimensions in delegated Prelu operators PiperOrigin-RevId: 317735911 Change-Id: Iddf727f5f916b142412a1be44efa1f367dc31d49
-
由 Geoffrey Martin-Noble 提交于
These differ from the CMake library names somewhat arbitrarily, which just creates confusion. PiperOrigin-RevId: 317735676 Change-Id: Ia9d3d1645467b1af72f917b01d1a3f6061bceb34
-
由 Scott Wegner 提交于
functions PiperOrigin-RevId: 317735476 Change-Id: I58123c5fb64257975e220df4e57f2a1f74856da7
-
由 TensorFlower Gardener 提交于
PiperOrigin-RevId: 317735359 Change-Id: Ia8fe2425b21411aa2d26f7728e3c70d550fecb9a
-
由 Rick Chao 提交于
MultiProcessRunner: Add UnexpectedSubprocessExitError to be raised if the exit code from a subprocess in unexpected. This results in subprocess having segfault failing the test, which would not have before this change. PiperOrigin-RevId: 317734369 Change-Id: I008f9bd4dd5b7a68e8e4a7f17b7e5d490a30c09a
-