update to latest #9
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
PiperOrigin-RevId: 320129620 Change-Id: I76c1f633c341ef931be8d3eb2d378ad3e7f12fda
…nce users are not expected to use them directly. PiperOrigin-RevId: 320129981 Change-Id: Ifd3ec638549de95c01147560260b8ecf7f17e29d
…egment_graph PiperOrigin-RevId: 320132418 Change-Id: Ia355393a792bdd0d60bf4d179c7b43c8f6bdcfa9
…ives. Prior to this change, TF_TString_Init intentionally only initialized the first 2 bytes of a 24 byte struct. In some situations, where a temporary TF_TString struct is initialized via TF_TString_Init and subsequently copied, the copying of the uninitialized data may result in an msan use-of-uninitialized-value error. PiperOrigin-RevId: 320133690 Change-Id: I995ba629c4cb0b2a85a25c277caa74cf2dbc3bc3
PiperOrigin-RevId: 320135405 Change-Id: I282170e72bfc74354d56f11c4e22d67886fc7296
PiperOrigin-RevId: 320138569 Change-Id: Ia0f4549d97b05a993eaf6aab7ce700f205106d5a
PiperOrigin-RevId: 320141724 Change-Id: Ic3033d235884d32cdf3c92df9068fb4415d45e53
…ersion. PiperOrigin-RevId: 320146856 Change-Id: Ic534e97b2eecbd4573b91ff48ef90d38bbacd9a4
PiperOrigin-RevId: 320148492 Change-Id: Ie8da7f170556efa6a973601300079f6916785101
PiperOrigin-RevId: 320148494 Change-Id: Id05e21e7aa78eaf55e9300b411c26fc0cb0573a8
PiperOrigin-RevId: 320149593 Change-Id: Ie85979a26225dbf017ad4bb888de1548e55a6d7f
This error is triggered when the XLA program outfeeds data with a shape that is not compatible with the shape expected by the runtime. This error can be triggered by the end user. Currently, the error is not very helpful. This change adds the two mismatched shapes to the error message, in the spirit of the error message reported on CPU for the same error. (see tensorflow/compiler/xla/service/cpu/cpu_runtime.cc line 289) PiperOrigin-RevId: 320162796 Change-Id: I723e70f740a2b0734b7be575ea20d5bed19f55e7
PiperOrigin-RevId: 320172723 Change-Id: Ib81a360aa8e85b778614e0a0e2aee5cf8947ad39
This is on-par with the existing emitter, sometimes better and unlocks more potential. The strategy classes are duplicated right now, but I expect them to graduate to mlir core soon. I'm planning to remove the custom LLVM IR emitters if this turns out to be stable enough. PiperOrigin-RevId: 320175958 Change-Id: I2df3d09c40041f8a69dac1cb45e945af203ec6e1
PiperOrigin-RevId: 320184343 Change-Id: I10d3778dbb0dc86450c6e1ad69cd35b9fdffe8a0
PiperOrigin-RevId: 320185037 Change-Id: Ia2149c852d6cf0a07d36c96aace3bf1c3fcb12b9
…mation in the graph nodes. Previously, the TF-TRT bridge may put the graph nodes that are intended to run on a non-GPU device to run with TensorRT. It may also put two graph nodes that are intended to run on two specific but different GPUs to one TensorRT engine which will make the two nodes run on the same GPUs. This change modifies the bridge to only put nodes that are not intended to run on non-GPU devices to run with TensorRT and to only group nodes with compatible device assignments into a clust for a TensorRT engine. Add tests to test cluster formation in the presence of device assignment. Modify an existing test to test conversion with device assignment. PiperOrigin-RevId: 320191854 Change-Id: I76aea67abc7634d73ba07bccec52ac25df65735b
PiperOrigin-RevId: 320193024 Change-Id: I788ebf903e14e9c0acd3eee58a86fb5afd14a16d
PiperOrigin-RevId: 320196263 Change-Id: I895b7295e698f7005b74ffc42bc04eb4a9a87b75
MLIR bridge doesn't support tf.NonMaxSuppressionV4 legalization that is conditionally generated by non_max_suppression_padded function. PiperOrigin-RevId: 320197235 Change-Id: If7242133254680b366771ced50de074ed6180563
Searches through gc.get_referrers for a few hops. Seems to find variables and tables, which will cover most cases. PiperOrigin-RevId: 320199554 Change-Id: I69e68ab86d064f689b2248b25fa486fb99096652
PiperOrigin-RevId: 320200195 Change-Id: Ifbdb7d418c117129e54a379be3e16b522ee920fe
…distribute. PiperOrigin-RevId: 320202511 Change-Id: I0c57c90847915bd702e4fa204653a76f7cd65c38
PiperOrigin-RevId: 320204496 Change-Id: I069dbb1a18eadde1f6eaa0d0d7e5c02d1799e07c
PiperOrigin-RevId: 320205862 Change-Id: Iad999a1a85d3d193258017d8a042973d57163492
- Also adopt variadic llvm::isa<> in more places PiperOrigin-RevId: 320206113 Change-Id: Ia03a1503f699fb6be6dff02e90b6630d6d894b19
The root cause is probably related to keras global graph and some potential memory leak. Adding a teardown method to force clean any keras model created in the test, with a force GC. Also cleanup some assertion methods which make the test code more readable. PiperOrigin-RevId: 320630906 Change-Id: Ib72a39c184613e431d6eab03b18f3b217cb8e506
Imported from GitHub PR #39429 This PR fixes and enables the TopK op and its tests for ROCm. Copybara import of the project: -- 7ef56e6 by Eugene Kuznetsov <eugene.kuznetsov@amd.com>: Fixing and enabling TopK on ROCm COPYBARA_INTEGRATE_REVIEW=#39429 from ekuznetsov139:google-upstream-topk 7ef56e6 PiperOrigin-RevId: 320631998 Change-Id: If5291d7ff5ff0e98f953645e0adf085bf891cee8
…TfLiteContext. TFLM wants to reduce RAM usage for model allocation. This is the first phase to reduce this overhead. Instead of directly allocating TfLiteTensor items onto TfLiteContext, TF Micro will return instances through function pointers on TfLiteContext. The new TfLiteEvalTensor will be used in TFLM kernel implementations to reduce the memory overhead during TfLiteRegistration::Eval() calls. NOTE: TfLiteEvalTensor can be moved into TFLM-only build rules when internal builds use TF_LITE_STATIC_MEMORY by default (b/160955687). Additionally, TfLiteContext contains many fields not used by TFLM and should be forked to reduce memory overhead as well. PiperOrigin-RevId: 320634141 Change-Id: I26d49dfd5fa8f96bea8e098202d191d7ae6f1957
PiperOrigin-RevId: 320634923 Change-Id: I3991f0a8567006f9312049c805537258bdb8b478
PiperOrigin-RevId: 320635155 Change-Id: I0c254059f2e04c69774680c4fc28bb690bf7d0f9
…ocm_platform_misc_200708 PiperOrigin-RevId: 320639190 Change-Id: Icbeca0caa8f7601eaf6ae07385f9d07bd597d72d
PiperOrigin-RevId: 320640683 Change-Id: Iaf606eb19770c2495cd7da4196566f986d584b67
…t they correctly trigger dispatching. Note: Although this changes the api_def pbtxt files to change how exporting happens, the actual exported apis should not changed. PiperOrigin-RevId: 320641446 Change-Id: I70760649794f08fffb629e25244802f73cf8b256
… convert it to the NMSv4 TFLite op if there are no batches. PiperOrigin-RevId: 320643078 Change-Id: I8afb0edd7e384bad36d686c910dd183c843678e5
- Eliminate custom checks for checking no region arguments in IfRegion verification. - update expected error messages in IfRegion unit tests. PiperOrigin-RevId: 320643657 Change-Id: I982ea71cf3bffe9811773724d75a79bb98cd35b7
PiperOrigin-RevId: 320644195 Change-Id: I4b3fc1c06e3fc9e9a5d5731f1100ae6ae801598c
PiperOrigin-RevId: 320645237 Change-Id: I08702860e9d71149a01565293dbdb1fc29443833
PiperOrigin-RevId: 320647339 Change-Id: Ibd925006f1a1abd5c09ea815ec9e6d62ac7982c6
PiperOrigin-RevId: 320648488 Change-Id: I719b3cd44392e7699f80c697c9c64a75cdf6ac15
PiperOrigin-RevId: 320648545 Change-Id: If31a8b3ae8598fbd06129f052211fbf8d4583990
PiperOrigin-RevId: 320654953 Change-Id: I765f140dd5ab88c34d916cd0fe429d50c56edd02
PiperOrigin-RevId: 320655274 Change-Id: Ie7c4bfa6bdef5120f8fe4afd4cc26cb84efa39c0
PiperOrigin-RevId: 320656309 Change-Id: I6f749752b9d29a3cb474d0e0930e8cbec293ce2d
PiperOrigin-RevId: 320661178
… experimental mlir bridge for TensorFlow V1 even when the user overrides Session's default ConfigProto. In TensorFlow V1, when the user doesn't override Session's default ConfigProto, the ConfigProto uses the value from TF_XLA_FLAGS (via context.config). However, when the user passes in their own ConfigProto, enable_mlir_bridge is defaulted to false. PiperOrigin-RevId: 320661288 Change-Id: I5deacb4e12b6551a57c2496c35830ac3af25fe28
PiperOrigin-RevId: 320661639 Change-Id: I430d485c50eb2d6bed0cd4f2c93b02f6f1b1908f
PiperOrigin-RevId: 320662277 Change-Id: I15b033ef3be5484bc6c68e115023b923a4691180
…on host. PiperOrigin-RevId: 320664808 Change-Id: I41dac6e90bbe998ee6815055dd4a5edc04854767
PiperOrigin-RevId: 320664868
Otherwise dynamic dimension inference won't have the latest view of the graph. PiperOrigin-RevId: 320667881 Change-Id: I75f8e993904385fc516f046c96343fe54419e27f
PiperOrigin-RevId: 320669068 Change-Id: I18e790f5e9036e98571edf3094a4c5b47cb87874
get_shape() is an alias, and not implemented on RaggedTensor. PiperOrigin-RevId: 320669349 Change-Id: I2c2c6d36894fe2c76d9492b73cae01d009a303fb
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.