- 19 12月, 2019 40 次提交
-
-
由 Rohan Jain 提交于
Handle ExternalStatePolicy during Iterator::Save(). This CL adds an optional external_state_policy attr to the SerializeIterator op which is by default set to FAIL thereby keeping the behavior the same as before. Now users can override it if they want something else. This CL also exposes this option in the CheckpointInputPipelineHook. PiperOrigin-RevId: 286314588 Change-Id: I4f8eae88f42c74b168348afca8b295eb37d845dc
-
由 Amit Patankar 提交于
PiperOrigin-RevId: 286313739 Change-Id: I9d4dfb66a8e82c56249bc7f7b70c71d228912e9f
-
由 Dan Moldovan 提交于
Verify the early stopping condition in for loops before consuming the target, rather than just at the beginning of the iteration. PiperOrigin-RevId: 286311186 Change-Id: Ie2afdd2a9f26d5045baf4517141517c0d28e3879
-
由 Hye Soo Yang 提交于
PiperOrigin-RevId: 286311054 Change-Id: I1225b7ed8f6ca83368dac2bb1572b6abc7945fd3
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 286310575 Change-Id: I5027b42f4a620c15420c591bcc4a0909466da7fc
-
由 Rohan Jain 提交于
In all cases, we can't rely on the tensor.device attribute being set. So its better to get the device for a SaveSpec from the device passed in rather. This was an issue with saving iterators because for iterators the resource usually has a device specification but the serialized tensor derived from it might not have it set. As a result, when saving iterators in a sharded fashion all iterators end up on '' device instead which is not what is intended. Also adding support for saving iterators in a sharded fashion to avoid unnecessary copying during checkpointing. PiperOrigin-RevId: 286310419 Change-Id: I1a957af783f7f69753992ce220b59eb43df2c02f
-
由 Brian Zhao 提交于
Adding a define for TENSORFLOW_LITE_PROTOS; this gets the android build working with the granular target form core/platform:env. PiperOrigin-RevId: 286308504 Change-Id: I6dd05fbcc62000836bc03866c07677cde33b1652
-
由 Juhyun Lee 提交于
PiperOrigin-RevId: 286308295 Change-Id: I3df054ffbab15a1db933c6242c68e7fbc0b415e5
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 286307709 Change-Id: Ib936279d64e8faf8e452b95a08a9c4a3c3026c4a
-
由 Pavithra Vijay 提交于
PiperOrigin-RevId: 286305667 Change-Id: I4b016ac41492d52f361a121f2c303cdd753fbae8
-
由 Jacques Pienaar 提交于
Use newly added LookUp call. PiperOrigin-RevId: 286305459 Change-Id: Ibb5294ce98904bac6abbd902b0109715fffa34be
-
由 Tiezhen WANG 提交于
PiperOrigin-RevId: 286305113 Change-Id: I43539ef03478ad19a8329362d74f02c6e6035db8
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 286304583 Change-Id: I2e3b112c8c4c0b8225fc5ade2ff48e6c83f40dfd
-
由 Andrew Audibert 提交于
The effect can be seen in the Args section of https://www.tensorflow.org/api_docs/python/tf/data/Dataset?version=nightly#shuffle PiperOrigin-RevId: 286304265 Change-Id: I318caf0b33a92d881ad42065d0e4a7a603d91fc0
-
由 Jose Baiocchi 提交于
PiperOrigin-RevId: 286302648 Change-Id: I81960fb8344aaa17430b4c14d2edbf16c175b052
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 286302514 Change-Id: I57482eb77ac83f71253fe6b7ed716a969e09cf5f
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 286302425 Change-Id: I9f4e7e7b4c2c12ce1abe2fc1b176dd418944d02b
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 286302404 Change-Id: I4357302a07f2dca5175d739c51d25972c12892ae
-
由 A. Unique TensorFlower 提交于
This is part of a larger effort to deprecate swig and eventually with modularization break pywrap_tensorflow into smaller components. Please refer to https://github.com/tensorflow/community/blob/master/rfcs/20190208-pybind11.md for more information. PiperOrigin-RevId: 286302183 Change-Id: I4baf4a2628d46d7bdf3aa2916fb6f980a3c99abe
-
由 Andy Ly 提交于
ConvertMaxPoolOp, ConvertRangeOp, and ConvertSigmoidOp were added to the patterns list twice. PiperOrigin-RevId: 286300831 Change-Id: I9106e8fa3c275b9ed78f5176715d0a803463f1e4
-
由 Jacques Pienaar 提交于
Ran into case where failed constant folds resulted in non-deterministic test behavior (e.g., failed to fold unrelated op). Need to dig into this a bit more. PiperOrigin-RevId: 286299132 Change-Id: I290dade5fba0fd5b08dae4d696878b9196fc516d
-
由 Lucy Fox 提交于
In the training case, the TF op lowers to HLO op BatchNormTraining. Otherwise, it lowers to HLO op BatchNormInference. PiperOrigin-RevId: 286297657 Change-Id: If8e14edcc8f43016f273ce29ef503ea83d639a25
-
由 Amit Patankar 提交于
This version fixed a segment fault issue in GRPC. See https://github.com/grpc/grpc/pull/19531. PiperOrigin-RevId: 286294784 Change-Id: I11c585833f952fac224990e4d3e2b350a3ba1416
-
由 Nupur Garg 提交于
PiperOrigin-RevId: 286290570 Change-Id: Ia30be150c3a25c81fc17a458fc3f4b85cc3da11f
-
由 Dong Lin 提交于
This version fixed a segment fault issue in GRPC. See https://github.com/grpc/grpc/pull/19531. PiperOrigin-RevId: 286289959 Change-Id: I7c107f8102b3a98b493b1992bec39a4794ad23c4
-
由 Zhenyu Tan 提交于
PiperOrigin-RevId: 286288622 Change-Id: I2625c877389c5a627d96ed82025d3a623c9bbb6f
-
由 Francois Chollet 提交于
PiperOrigin-RevId: 286287077 Change-Id: Iea055499b8900fdb4903b7a22ca6a091e548bffe
-
由 Yunlu Li 提交于
PiperOrigin-RevId: 286286868 Change-Id: I92bb40fb8eff9a16e5b382fec19befd39c93aeaa
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 286283327 Change-Id: I8fd9102c04ffe56af92fa19b01bcece139c45bae
-
由 Frank Chen 提交于
PiperOrigin-RevId: 286281695 Change-Id: I021eec024331e65b2d6243e14bb477957ce0e2f6
-
由 Frank Chen 提交于
PiperOrigin-RevId: 286281590 Change-Id: I2c87462702d3ec9b6e92a23968f5b02745adb777
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 286279090 Change-Id: I5f817cd51b8516ff581093413992b199674d5bdd
-
由 Nick Kreeger 提交于
This version varies slightly from the current reference implementation in TFLite (original): 1.) All references to tensor_utils:: namespace are dropped due to the build incompatibility and size the include brings in (just like the rest of this port for float/hybrid-quant). 2.) Scratch tensors are re-worked into variable tensors. This is a temporary workaround until memory planning lands . 3.) An additional Tensor is required to provide pre-calculated scale values. These calculations are very expensive on low power device. PiperOrigin-RevId: 286278125 Change-Id: Ibbadb2f38a6c25b5550b4fced5b32c7a5b9420df
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 286277617 Change-Id: Ief36745883c36d48a0aad4f49aee806573502cc1
-
由 Jose Baiocchi 提交于
PiperOrigin-RevId: 286276718 Change-Id: Ide2445cc395c1a28c4faf176034c2f34f41cc50a
-
由 Jose Baiocchi 提交于
PiperOrigin-RevId: 286276313 Change-Id: I0010b5cceb096b639f8a8cde82debfcbae177d51
-
由 Amit Patankar 提交于
PiperOrigin-RevId: 286275744 Change-Id: Ie82e11c606bb65a0853c1432d1336cca290b636c
-
由 Yifei Feng 提交于
Looks like https://launchpad.net/~jonathonf is getting slowly removed. Fix "The repository 'http://ppa.launchpad.net/jonathonf/python-3.6/ubuntu xenial Release' does not have a Release file." PiperOrigin-RevId: 286273233 Change-Id: I5fd9b896f03f748fea440d5ab4077d996ebe2afd
-
由 Yash Katariya 提交于
PiperOrigin-RevId: 286271783 Change-Id: I6acc71de279834957bd22271990609f79e1e2517
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 286257478 Change-Id: I56574945f4f0aba418e2bfb8dbe3a510713a65f4
-