aboutsummaryrefslogtreecommitdiffhomepage
path: root/tensorflow/contrib/lite/toco
Commit message (Collapse)AuthorAge
* Move tflite_convert g3docs, so they will be pulled into the site.Gravatar Mark Daoust2018-10-09
| | | | PiperOrigin-RevId: 216452447
* Return ::tensorflow::Status in Toco Graph Transformations.Gravatar Yu-Cheng Ling2018-10-09
| | | | PiperOrigin-RevId: 216392908
* Update TFLite Converter documentation.Gravatar Nupur Garg2018-10-09
| | | | PiperOrigin-RevId: 216386450
* Fixing Toco for exporting graphs with stringsGravatar A. Unique TensorFlower2018-10-09
| | | | | | | If the graph contains not constant array with strings it fails because the array's size can't be estimated. PiperOrigin-RevId: 216356162
* Fix issue with type inference for ops with fixed output typesGravatar Jared Duke2018-10-08
| | | | | | | Use the ArgDef::type field when available for propagating the output types from a given unsupported operator. PiperOrigin-RevId: 216257741
* Remove the restrictions that constant resolution of reduce_sum operators ↵Gravatar A. Unique TensorFlower2018-10-08
| | | | | | must be on axis 0, and can only be on 1 or 2-d inputs. PiperOrigin-RevId: 216226776
* Handle Range & BatchMatMul in partial Flex modeGravatar Yu-Cheng Ling2018-10-05
| | | | PiperOrigin-RevId: 215957535
* Refactoring TFLite export code. Unify OperatorCode generation logic.Gravatar Yu-Cheng Ling2018-10-05
| | | | PiperOrigin-RevId: 215928419
* Automated rollback of commit 8dc7bc7764150253c03a666eee84fc48f867d6a2Gravatar A. Unique TensorFlower2018-10-04
| | | | PiperOrigin-RevId: 215768310
* TFLite Flex: Blacklist Control Flow OpsGravatar Yu-Cheng Ling2018-10-03
| | | | PiperOrigin-RevId: 215658384
* In all constant-propagation transformations, check that the array we'd be ↵Gravatar A. Unique TensorFlower2018-10-02
| | | | | | | | | | | | | | | | | 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 infoGravatar A. Unique TensorFlower2018-10-02
| | | | PiperOrigin-RevId: 215487633
* Fix a bug: the use of sequence-point boolean operators here had theGravatar A. Unique TensorFlower2018-10-02
| | | | | | | 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 elementsGravatar A. Unique TensorFlower2018-10-02
| | | | | | | | 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
* Allow passing --allow_nonexistent_arrays via toco_convertGravatar A. Unique TensorFlower2018-10-02
| | | | PiperOrigin-RevId: 215440829
* Generate an error when --rnn_states refers to array names that aren't ↵Gravatar A. Unique TensorFlower2018-10-02
| | | | | | produced/consumed by any op. PiperOrigin-RevId: 215402308
* Rename TFLite Eager delegate -> Flex delegateGravatar Yu-Cheng Ling2018-09-27
| | | | PiperOrigin-RevId: 214835588
* Reduce the size of //tensorflow/tools/pip_package:simple_console_windowsGravatar A. Unique TensorFlower2018-09-27
| | | | | | | | This change reduce the size of //tensorflow/tools/pip_package:simple_console_windows's zip file from 1000027677 bytes to 47690474 bytes for a CPU build. For GPU build, it will avoid going over 4GB when multiple CUDA compatibility are specified. To fix #22390 PiperOrigin-RevId: 214764423
* Rename TocoConverter to TFLiteConverter.Gravatar Nupur Garg2018-09-26
| | | | PiperOrigin-RevId: 214710175
* Automated rollback of commit 82af048bc8c3c044c98a27b1c4c27bb62d4e4a14Gravatar Nupur Garg2018-09-26
| | | | PiperOrigin-RevId: 214705311
* Rename TFLite Eager delegate -> Flex delegateGravatar Yu-Cheng Ling2018-09-26
| | | | PiperOrigin-RevId: 214674717
* Allow empty arrays to occur as the first input to the concat op.Gravatar A. Unique TensorFlower2018-09-25
| | | | | | | The conversion process fails for graphs that use tf.boolean_mask(..., axis=0) -- this op calls tf.concat with an empty array as the first argument. PiperOrigin-RevId: 214451470
* Updated to newest FlatBuffers and FlexBuffers.Gravatar A. Unique TensorFlower2018-09-24
| | | | PiperOrigin-RevId: 214309210
* Update links to install pages.Gravatar Billy Lamberta2018-09-20
| | | | PiperOrigin-RevId: 213917946
* Fix missing TODO.Gravatar Suharsh Sivakumar2018-09-20
| | | | PiperOrigin-RevId: 213885561
* TOCO transformations updated to support dilated depthwise convolution.Gravatar Suharsh Sivakumar2018-09-19
| | | | PiperOrigin-RevId: 213729750
* Two improvements in resolve_tensorflow_matmul:Gravatar A. Unique TensorFlower2018-09-19
| | | | | | | | | | | | | | | | | | | | 1. Before inserting a new Transpose node, check if there already is one that may be reused. In practice, there are two cases: either the array being transposed is a constant (by far the most common case) or it's not. * If it is constant, then this doesn't really make a difference: ResolveConstantTranspose runs anyway, eliminating these Transpose nodes and also mootifying this change as it leaves no Transpose node to be reused. So in that case, constant-array-deduping is really the only thing that prevents duplication of data. * If it is not constant, that's where this new logic really helps, as the resulting Transpose nodes are here to stay in the final graph, and this avoids inserting more than are needed. 2. transpose_a is not supported. However, rather than CHECK-fail, it's more useful to have this graph transformation bail with a log message. The resulting 'unresolved' MatMul node could still be handled in some way at the TFLite level, or we could end up having support for MatMul per se. PiperOrigin-RevId: 213678294
* Fixes in ResolveReorderAxes.Gravatar A. Unique TensorFlower2018-09-19
| | | | | | The main issue is we were keeping the input array, updating it in place and discarding the output array. That was a problem when the input array had multiple consumer ops. Now we're keeping the output array instead, which is the correct thing to do. However, in order to minimize disruption, we keep using the input array's name whenever possible, by means of some array renamings. PiperOrigin-RevId: 213678219
* Enable large constant array deduping by default.Gravatar A. Unique TensorFlower2018-09-19
| | | | | | | If this causes trouble (makes graph visualizations harder to read, etc) then consider increasing the default value of dedupe_array_min_size_bytes. PiperOrigin-RevId: 213656796
* Implement ZerosLikeGravatar A. Unique TensorFlower2018-09-17
| | | | PiperOrigin-RevId: 213227615
* Improve output parsing for unsupported opsGravatar Jared Duke2018-09-14
| | | | PiperOrigin-RevId: 213017532
* Reland "Add basic type propagation for unsupported ops in TFLite conversion"Gravatar Jared Duke2018-09-13
| | | | | | | The original CL was rolled back due to op registration conflicts in the pip. Resolve the issue by only including core:ops in the toco binary itself, not in intermediate libraries. PiperOrigin-RevId: 212902838
* Automated rollback of commit 6b507a6de855a6f988100904229b7f46a5652b88Gravatar Suharsh Sivakumar2018-09-13
| | | | PiperOrigin-RevId: 212890622
* Dilated Depthwise Conv reference implementations.Gravatar Suharsh Sivakumar2018-09-13
| | | | PiperOrigin-RevId: 212884951
* Add basic type propagation for unsupported ops in TFLite conversionGravatar Jared Duke2018-09-12
| | | | PiperOrigin-RevId: 212651704
* Implementation of square.Gravatar A. Unique TensorFlower2018-09-11
| | | | PiperOrigin-RevId: 212577288
* [TF] Variant improvements.Gravatar Eugene Brevdo2018-09-11
| | | | | | | | | | | | | | | | | | 1. Change Variant Decode to accept VariantTensorData (non-ref). This should allow some optimization in the future. In the meantime it means removing the variant.h include from tensor.h, since variant_encode_decode.h now relies on tensor.h and variant.h now relies on that. It also means we found a bunch of places where tensor.proto.h, variant.h, and mutex.h were being imported through tensor.h (along with a bunch of other crap); so now we directly import them in order to compile. 2. Move Variant registry to use TypeIndex instead of a TypeName string; this should speed up registry lookups. PiperOrigin-RevId: 212478896
* Merge pull request #21611 from kouml:removing_redundant_semicolonGravatar TensorFlower Gardener2018-09-10
|\ | | | | | | PiperOrigin-RevId: 212346420
* | Convert more kernel signatures to use runtime shapes.Gravatar A. Unique TensorFlower2018-09-06
| | | | | | | | PiperOrigin-RevId: 211874785
* | Update diagram in TOCO README.Gravatar Nupur Garg2018-09-05
| | | | | | | | PiperOrigin-RevId: 211711493
* | Minimum change for generating Eager ops with Toco.Gravatar Yu-Cheng Ling2018-09-05
| | | | | | | | PiperOrigin-RevId: 211621189
* | Hardcode input range from output for reluGravatar A. Unique TensorFlower2018-09-04
| | | | | | | | PiperOrigin-RevId: 211562900
* | Minor cleanup: Simplify declaration of transformation.Gravatar Suharsh Sivakumar2018-08-31
| | | | | | | | PiperOrigin-RevId: 211175130
* | Fixing Any operator.Gravatar A. Unique TensorFlower2018-08-31
| | | | | | | | PiperOrigin-RevId: 211159438
* | Introduce post_training_quantize flag and deprecate quantize_weights flag.Gravatar Suharsh Sivakumar2018-08-31
| | | | | | | | PiperOrigin-RevId: 211124183
* | Documentation fixes.Gravatar Nupur Garg2018-08-31
| | | | | | | | PiperOrigin-RevId: 211092680
* | Call new tflite buffer quantize weights transformation tool from TOCO.Gravatar Suharsh Sivakumar2018-08-30
| | | | | | | | PiperOrigin-RevId: 211020126
* | Enable quantization support for TopKGravatar Ben Lee2018-08-30
| | | | | | | | PiperOrigin-RevId: 211011100
* | Change the data type of mean_values and std_dev_values to float.Gravatar Yunlu Li2018-08-30
| | | | | | | | PiperOrigin-RevId: 211010293
* | Adding support for resolving constant Exp ops.Gravatar A. Unique TensorFlower2018-08-30
| | | | | | | | PiperOrigin-RevId: 210995878
generated by cgit v1.2.3 (git 2.39.1) at 2024-06-24 14:39:18 +0000