Commit message (Collapse) | Author | Age | |
---|---|---|---|
* | Use Ophints to support TfLite UnidirectionaSequenceLstm and add an e2e test. | A. Unique TensorFlower | 2018-10-09 |
| | | | | | | Support peephole and num_proj as well. PiperOrigin-RevId: 216467578 | ||
* | 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 | ||
* | Allow passing --allow_nonexistent_arrays via toco_convert | A. Unique TensorFlower | 2018-10-02 |
| | | | | PiperOrigin-RevId: 215440829 | ||
* | Generate an error when --rnn_states refers to array names that aren't ↵ | A. Unique TensorFlower | 2018-10-02 |
| | | | | | | produced/consumed by any op. PiperOrigin-RevId: 215402308 | ||
* | Implement ZerosLike | A. Unique TensorFlower | 2018-09-17 |
| | | | | PiperOrigin-RevId: 213227615 | ||
* | Implementation of unpack op. | A. Unique TensorFlower | 2018-08-23 |
| | | | | PiperOrigin-RevId: 210051131 | ||
* | Propagate narrow_range flags in suitable constant-propagation | A. Unique TensorFlower | 2018-08-17 |
| | | | | | | transformations. PiperOrigin-RevId: 209174233 | ||
* | Allow empty shapes in certain cases within toco | Jared Duke | 2018-08-08 |
| | | | | PiperOrigin-RevId: 207913842 | ||
* | Implementation of ctc beam search decoder op in custom op fashion. | A. Unique TensorFlower | 2018-08-02 |
| | | | | PiperOrigin-RevId: 207210333 | ||
* | More explicit TOCO error message when the input graph is missing min/max needed | A. Unique TensorFlower | 2018-07-30 |
| | | | | | | for quantization. PiperOrigin-RevId: 206650116 | ||
* | Implementation of logical_or. | A. Unique TensorFlower | 2018-07-30 |
| | | | | PiperOrigin-RevId: 206549781 | ||
* | Add one_hot op support to TFLite | Jared Duke | 2018-07-26 |
| | | | | PiperOrigin-RevId: 206185190 | ||
* | Allow scalar inputs for toco and TFLite zip tests | Jared Duke | 2018-07-23 |
| | | | | PiperOrigin-RevId: 205674334 | ||
* | Import/export support for Any, LogicalAnd, and LogicalNot ops. | A. Unique TensorFlower | 2018-07-18 |
| | | | | PiperOrigin-RevId: 205134621 | ||
* | Fixing bugs in graphdef export of bool types and a few broken ops. | A. Unique TensorFlower | 2018-07-18 |
| | | | | PiperOrigin-RevId: 205111208 | ||
* | toco thinks Stack is the same as Pack. | A. Unique TensorFlower | 2018-07-17 |
| | | | | | | | | | | | | | | Stack is not Pack. tf.stack() yields Pack ops. Pack ops stack tensors. Stack ops manipulate the TF runtime stack. This cl unaliases "Stack" and "Pack" ops in toco, and renames most things that refer to "Stack" ops to "Pack" to be consistent across the codebase. In summary: Stack is whack. 'Stack' should be 'Pack'. Hack 'Stack's into 'Pack's like a maniac. This keeps 'Stack's from wracking runtime graphs. (We apologize for the fault in the change description. Those responsible have been...er...sacked). PiperOrigin-RevId: 204951155 | ||
* | Support reduce_max and reduce_prod | A. Unique TensorFlower | 2018-07-16 |
| | | | | PiperOrigin-RevId: 204846139 | ||
* | Implementation of arg_min. | A. Unique TensorFlower | 2018-07-10 |
| | | | | PiperOrigin-RevId: 203908601 | ||
* | Update reshape shape field when introducing a copy. | A. Unique TensorFlower | 2018-07-07 |
| | | | | PiperOrigin-RevId: 203522934 | ||
* | Fix array logging bug | A. Unique TensorFlower | 2018-07-02 |
| | | | | PiperOrigin-RevId: 203021167 | ||
* | Implementation of pow. | A. Unique TensorFlower | 2018-06-27 |
| | | | | PiperOrigin-RevId: 202262513 | ||
* | Un-fused quantized Babelfish LSTM cell support in TFLite | A. Unique TensorFlower | 2018-06-26 |
| | | | | | | including support for shuffled-weights fully-connected op. PiperOrigin-RevId: 202192299 | ||
* | Make fused activation opt-in | Jared Duke | 2018-06-20 |
| | | | | PiperOrigin-RevId: 201450857 | ||
* | Fix operator names. | A. Unique TensorFlower | 2018-06-20 |
| | | | | PiperOrigin-RevId: 201422566 | ||
* | Fix for DumpGraphviz() failing on zero-sized arrays, which it should handle ↵ | A. Unique TensorFlower | 2018-06-13 |
| | | | | | | for debugging purposes. PiperOrigin-RevId: 200397151 | ||
* | Leverage the standard error space by using tensorflow::Status | A. Unique TensorFlower | 2018-06-12 |
| | | | | PiperOrigin-RevId: 200322035 | ||
* | Automated g4 rollback of changelist 200228895 | Nupur Garg | 2018-06-12 |
| | | | | PiperOrigin-RevId: 200279737 | ||
* | Add strings type to TOCO Python API. | Nupur Garg | 2018-06-12 |
| | | | | PiperOrigin-RevId: 200228895 | ||
* | Optimizing transpose_conv. | A. Unique TensorFlower | 2018-06-08 |
| | | | | PiperOrigin-RevId: 199839745 | ||
* | Implementation of TensorFlowEqual and TensorFlowNotEqual. | A. Unique TensorFlower | 2018-06-07 |
| | | | | PiperOrigin-RevId: 199602232 | ||
* | implementation of sparse_to_dense | A. Unique TensorFlower | 2018-05-31 |
| | | | | PiperOrigin-RevId: 198710452 | ||
* | Various ClangTidy-inspired fixes. | A. Unique TensorFlower | 2018-05-14 |
| | | | | PiperOrigin-RevId: 196556727 | ||
* | Don't check that bool arrays are quantized. | A. Unique TensorFlower | 2018-05-14 |
| | | | | PiperOrigin-RevId: 196541955 | ||
* | Implement sin operator | A. Unique TensorFlower | 2018-05-09 |
| | | | | PiperOrigin-RevId: 196062186 | ||
* | Add support for select (via tf.where) support to tflite. | A. Unique TensorFlower | 2018-05-07 |
| | | | | PiperOrigin-RevId: 195734246 | ||
* | add support for PadV2 | A. Unique TensorFlower | 2018-05-05 |
| | | | | PiperOrigin-RevId: 195503894 | ||
* | Some fixes to support another TF graph: | A. Unique TensorFlower | 2018-05-04 |
| | | | | | | | | | | | | | 1. Fix ResolveBatchNormalization to avoid deleting arrays that may still be used. 2. Correctly count the number of ops using a given array, even when some ops use the same array as more than one of their inputs. 3. In PropagateFixedSizes for Concatenation ops, when resolving a -1 wildcard to a fixed value, we were doing so in a local 'axis' variable without actually updating op->axis! The resulting -1 value still in op->axis tripped runtime code, causing the concatenation to misbehave during inference. PiperOrigin-RevId: 195454037 | ||
* | Post-transform pass to dedupe large constant arrays. | A. Unique TensorFlower | 2018-05-03 |
| | | | | PiperOrigin-RevId: 195260578 | ||
* | Automated g4 rollback of changelist 195091587 | A. Unique TensorFlower | 2018-05-02 |
| | | | | PiperOrigin-RevId: 195184798 | ||
* | Automated g4 rollback of changelist 195091587 | A. Unique TensorFlower | 2018-05-02 |
| | | | | PiperOrigin-RevId: 195098224 | ||
* | ArraysExtraInfo: Add name_regexp field and regexp name matching. | A. Unique TensorFlower | 2018-05-02 |
| | | | | PiperOrigin-RevId: 195091587 | ||
* | Merge changes from github. | Patrick Nguyen | 2018-05-01 |
| | | | | PiperOrigin-RevId: 194997009 | ||
* | Add fast shuffled fully-connected path also for the case where the | A. Unique TensorFlower | 2018-04-23 |
| | | | | | | | | | | | | | RHS has 4 columns (so far was only for the GEMV case where RHS has 1 column). Also pre-shuffle / pre-xor the input activations, not just the weights. We need a buffer for that, so the shuffled FullyConnected operator gets a second output acting as its workspace, similar to what we have been doing for Conv operators needed a im2col workspace buffer. PiperOrigin-RevId: 193958461 | ||
* | Allow default min/max ranges for int16 data types. | A. Unique TensorFlower | 2018-04-18 |
| | | | | PiperOrigin-RevId: 193362891 | ||
* | Fixes and cleanup to support more complex quantized models and adds ↵ | A. Unique TensorFlower | 2018-04-17 |
| | | | | | | PropagateFakeQuantNumBits. PiperOrigin-RevId: 193232630 | ||
* | ResolveConstantReshape transformation and fix for ResolveConstantTranspose. | A. Unique TensorFlower | 2018-04-12 |
| | | | | PiperOrigin-RevId: 192670991 | ||
* | Adding support for the standalone log operator. Basic import/export only, No ↵ | A. Unique TensorFlower | 2018-04-09 |
| | | | | | | run time support. PiperOrigin-RevId: 192135843 | ||
* | Add a command line parameter to toco to change the way toco rescales input ↵ | A. Unique TensorFlower | 2018-04-05 |
| | | | | | | and output tensors. PiperOrigin-RevId: 191825756 | ||
* | quantized LSTM support improvements | A. Unique TensorFlower | 2018-04-05 |
| | | | | PiperOrigin-RevId: 191794956 | ||
* | Permit use of ArraysExtraInfo/constant_float_value when target model is ↵ | A. Unique TensorFlower | 2018-04-04 |
| | | | | | | quantized. PiperOrigin-RevId: 191639289 |