Commit message (Collapse) | Author | Age | |
---|---|---|---|
* | Move tflite_convert g3docs, so they will be pulled into the site. | Mark Daoust | 2018-10-09 |
| | | | | PiperOrigin-RevId: 216452447 | ||
* | Adds an Objective-C API to TensorFlow Lite experimental. | A. Unique TensorFlower | 2018-10-09 |
| | | | | PiperOrigin-RevId: 216451263 | ||
* | Internal change. | A. Unique TensorFlower | 2018-10-09 |
| | | | | PiperOrigin-RevId: 216447412 | ||
* | Make lite_test.py run in open source. | Nupur Garg | 2018-10-09 |
| | | | | PiperOrigin-RevId: 216445964 | ||
* | Internal change | Jared Duke | 2018-10-09 |
| | | | | PiperOrigin-RevId: 216442906 | ||
* | Add support for time-major input in the bidirectional RNN Op. | A. Unique TensorFlower | 2018-10-09 |
| | | | | PiperOrigin-RevId: 216419983 | ||
* | Internal Change | A. Unique TensorFlower | 2018-10-09 |
| | | | | PiperOrigin-RevId: 216419037 | ||
* | Return ::tensorflow::Status in Toco Graph Transformations. | Yu-Cheng Ling | 2018-10-09 |
| | | | | PiperOrigin-RevId: 216392908 | ||
* | Update TFLite Converter documentation. | Nupur Garg | 2018-10-09 |
| | | | | PiperOrigin-RevId: 216386450 | ||
* | Internal change. | Nupur Garg | 2018-10-09 |
| | | | | PiperOrigin-RevId: 216385202 | ||
* | Internal change | Jared Duke | 2018-10-09 |
| | | | | PiperOrigin-RevId: 216367867 | ||
* | Fixing Toco for exporting graphs with strings | A. Unique TensorFlower | 2018-10-09 |
| | | | | | | | If the graph contains not constant array with strings it fails because the array's size can't be estimated. PiperOrigin-RevId: 216356162 | ||
* | Add Floor_mod to schema. | A. Unique TensorFlower | 2018-10-08 |
| | | | | PiperOrigin-RevId: 216303340 | ||
* | Internal change | Jared Duke | 2018-10-08 |
| | | | | PiperOrigin-RevId: 216270385 | ||
* | Fix issue with type inference for ops with fixed output types | Jared Duke | 2018-10-08 |
| | | | | | | | Use the ArgDef::type field when available for propagating the output types from a given unsupported operator. PiperOrigin-RevId: 216257741 | ||
* | Update performance documentation. | Shashi Shekhar | 2018-10-08 |
| | | | | PiperOrigin-RevId: 216248418 | ||
* | Remove the restrictions that constant resolution of reduce_sum operators ↵ | A. Unique TensorFlower | 2018-10-08 |
| | | | | | | must be on axis 0, and can only be on 1 or 2-d inputs. PiperOrigin-RevId: 216226776 | ||
* | Remove Dims from types.h, create build structure. | A. Unique TensorFlower | 2018-10-08 |
| | | | | PiperOrigin-RevId: 216191084 | ||
* | Add SequenceLSTMOptions to schema to decouple the sequential Op from the LSTM. | A. Unique TensorFlower | 2018-10-06 |
| | | | | PiperOrigin-RevId: 216066634 | ||
* | Internal change | Jared Duke | 2018-10-05 |
| | | | | PiperOrigin-RevId: 215978771 | ||
* | Add deprecation call-out for tf_mobile | A. Unique TensorFlower | 2018-10-05 |
| | | | | PiperOrigin-RevId: 215971335 | ||
* | Handle Range & BatchMatMul in partial Flex mode | Yu-Cheng Ling | 2018-10-05 |
| | | | | PiperOrigin-RevId: 215957535 | ||
* | Internal change. | A. Unique TensorFlower | 2018-10-05 |
| | | | | PiperOrigin-RevId: 215951354 | ||
* | Disable micro/examples/micro_speech:micro_speech_test test under msan | Smit Hinsu | 2018-10-05 |
| | | | | PiperOrigin-RevId: 215939542 | ||
* | Refactoring TFLite export code. Unify OperatorCode generation logic. | Yu-Cheng Ling | 2018-10-05 |
| | | | | PiperOrigin-RevId: 215928419 | ||
* | Relax some unnecessary 4D array restrictions | A. Unique TensorFlower | 2018-10-05 |
| | | | | PiperOrigin-RevId: 215910400 | ||
* | Merge the different LSTM EvalFloat/EvalHybrid calls into a single file. | A. Unique TensorFlower | 2018-10-05 |
| | | | | PiperOrigin-RevId: 215870962 | ||
* | Fix for memory issue in micro test code, spotted by asan checks | Pete Warden | 2018-10-04 |
| | | | | PiperOrigin-RevId: 215813259 | ||
* | Internal change. | A. Unique TensorFlower | 2018-10-04 |
| | | | | PiperOrigin-RevId: 215806953 | ||
* | Fix problem in quantized version of Comparison op handler | A. Unique TensorFlower | 2018-10-04 |
| | | | | PiperOrigin-RevId: 215801773 | ||
* | Use weak symbols to inject flex delegates | Jared Duke | 2018-10-04 |
| | | | | PiperOrigin-RevId: 215788183 | ||
* | Automated rollback of commit 8dc7bc7764150253c03a666eee84fc48f867d6a2 | A. Unique TensorFlower | 2018-10-04 |
| | | | | PiperOrigin-RevId: 215768310 | ||
* | Fix quantization util test to pass with defined behavior on 32-bit ↵ | Alan Chiao | 2018-10-04 |
| | | | | | | architectures. PiperOrigin-RevId: 215757844 | ||
* | Experimental interpreter, kernels, and example running TensorFlow Lite on a ↵ | Pete Warden | 2018-10-04 |
| | | | | | | microcontroller PiperOrigin-RevId: 215748973 | ||
* | Sparse output fully connected custom op. | Alan Chiao | 2018-10-04 |
| | | | | PiperOrigin-RevId: 215741296 | ||
* | Fix a test. | Shashi Shekhar | 2018-10-03 |
| | | | | | | | - SetCustomOp also sets the name of the custom op. Test was checking against the wrong name in the profile. PiperOrigin-RevId: 215665359 | ||
* | Automated rollback of commit c1b3b0b9e041d82e80c2cdcc623a387753daf0b4 | Anna R | 2018-10-03 |
| | | | | PiperOrigin-RevId: 215658770 | ||
* | TFLite Flex: Blacklist Control Flow Ops | Yu-Cheng Ling | 2018-10-03 |
| | | | | PiperOrigin-RevId: 215658384 | ||
* | Increase error-epsilon for ProfilingTest::ProfilesAreCollected. | Todd Wang | 2018-10-03 |
| | | | | PiperOrigin-RevId: 215654327 | ||
* | Fix 1970s-style bug in LogSoftmax eval. | A. Unique TensorFlower | 2018-10-03 |
| | | | | PiperOrigin-RevId: 215653797 | ||
* | Internal change. | Anna R | 2018-10-03 |
| | | | | PiperOrigin-RevId: 215631612 | ||
* | Add NNAPI padding enums to NeuralNetworksShim.h | A. Unique TensorFlower | 2018-10-03 |
| | | | | PiperOrigin-RevId: 215628561 | ||
* | Add the option of merging bidirectional RNN and LSTM outputs into a single ↵ | A. Unique TensorFlower | 2018-10-03 |
| | | | | | | | | output tensor. This is useful if the output of both directions will be passed to the next layer as a single output, as it avoids adding a concatenation op, which can be expensive on mobile devices where memory movement is relatively expensive. PiperOrigin-RevId: 215616140 | ||
* | Internal change. | Nupur Garg | 2018-10-03 |
| | | | | PiperOrigin-RevId: 215589009 | ||
* | Fix TfLiteTensor invalidation issue when using the Java API | Jared Duke | 2018-10-03 |
| | | | | | | | | Fix an issue where the Java Tensor class would hold a reference to an invalidated TfLiteTensor instance. This issue was manifest in certain models that add temporary tensors during execution. PiperOrigin-RevId: 215582842 | ||
* | Fix test that was relying on old lax toco behavior | A. Unique TensorFlower | 2018-10-03 |
| | | | | PiperOrigin-RevId: 215553161 | ||
* | In all constant-propagation transformations, check that the array we'd be ↵ | A. Unique TensorFlower | 2018-10-02 |
| | | | | | | | | | | | | | | | | | turning into a constant is a discardable array. If it's not discardable, it means that the user wants this array to keep existing in a way that is observable to them, i.e. not as weights. Typical example: a Fill op outputs an array that is passed as a RNN state array (non-discardable). It seems that so far we have been relying on accidental ordering of graph transformations for such state arrays not to be accidentally turned into constants. Instead, the desired graph transformation here is RemoveUnusedOp noticing that such a Fill can be discarded since its output is a RNN state array. So I don't have a test for this, but this seems to be tightening existing behavior, and should be good to have as long as it does not regress anything. PiperOrigin-RevId: 215500760 | ||
* | Make RemoveTrivialPassthrough preserve minmax-related info | A. Unique TensorFlower | 2018-10-02 |
| | | | | PiperOrigin-RevId: 215487633 | ||
* | Fix a bug: the use of sequence-point boolean operators here had the | A. Unique TensorFlower | 2018-10-02 |
| | | | | | | | unintended effect of causing the second line not to run at all depending on the result from the first line. PiperOrigin-RevId: 215466006 | ||
* | Do not warn about loss of accuracy in trivial cases when all array elements | A. Unique TensorFlower | 2018-10-02 |
| | | | | | | | | are equal to either the min or the max value, so that they are trivially exactly quantized. This case does not normally occur for true learned weights, which is what this warning is intended for. PiperOrigin-RevId: 215463096 |