Commit message (Collapse) | Author | Age | |
---|---|---|---|
* | Migrate a few conv kernels to use new kernel signatures. | A. Unique TensorFlower | 2018-09-27 |
| | | | | PiperOrigin-RevId: 214831837 | ||
* | Update kernel evals to use new kernel signatures. | A. Unique TensorFlower | 2018-09-24 |
| | | | | PiperOrigin-RevId: 214377809 | ||
* | Modularize TF Lite interface definitions and reorganize file structure | Pete Warden | 2018-09-07 |
| | | | | PiperOrigin-RevId: 212064501 | ||
* | Fixes to hybrid conv. Add additional tests for pointwise conv. | Alan Chiao | 2018-08-31 |
| | | | | PiperOrigin-RevId: 211085787 | ||
* | Optimized hybrid convolution with symmetric quantization. | Alan Chiao | 2018-08-28 |
| | | | | | | Add unit tests for multiple channels. PiperOrigin-RevId: 210567300 | ||
* | Quantized Dilated Convolution support. | Suharsh Sivakumar | 2018-08-14 |
| | | | | | | Also add tests for float and quantized dilated support, which seemed to be missing. PiperOrigin-RevId: 208766032 | ||
* | Generalize the Conv() op to support input product scale > output scale, i.e. | A. Unique TensorFlower | 2018-08-01 |
| | | | | | | | | | output_multiplier > 1. #20451 #19607 PiperOrigin-RevId: 206983654 | ||
* | Remove unnecessary <unistd.h> includes | Jared Duke | 2018-07-23 |
| | | | | PiperOrigin-RevId: 205744600 | ||
* | Make multithreaded conv respect setNumThreads() | A. Unique TensorFlower | 2018-07-07 |
| | | | | PiperOrigin-RevId: 203527657 | ||
* | Use persistent tensor for Convolution HWCN weights | Yu-Cheng Ling | 2018-06-28 |
| | | | | PiperOrigin-RevId: 202400843 | ||
* | add int32 support for add | A. Unique TensorFlower | 2018-06-28 |
| | | | | PiperOrigin-RevId: 202259189 | ||
* | Changed some variable names from camel case to underscore for consistency. | A. Unique TensorFlower | 2018-06-21 |
| | | | | PiperOrigin-RevId: 201587899 | ||
* | Standardize shifts in (more) multiplication util functions. | A. Unique TensorFlower | 2018-06-12 |
| | | | | PiperOrigin-RevId: 200271078 | ||
* | Bugfix for dilated_conv optimizations. We were failing to create im2col ↵ | A. Unique TensorFlower | 2018-06-08 |
| | | | | | | arrays for dilated unstrided 1x1 cases. PiperOrigin-RevId: 199849200 | ||
* | Make depthwiseconv handler handle filter ranges beyond 255 | A. Unique TensorFlower | 2018-05-23 |
| | | | | PiperOrigin-RevId: 197810361 | ||
* | Enhance error reporting. | A. Unique TensorFlower | 2018-05-21 |
| | | | | PiperOrigin-RevId: 197390052 | ||
* | Support 1x1x1xN bias sizes in TFLite's convolution and FC layers. | A. Unique TensorFlower | 2018-05-17 |
| | | | | PiperOrigin-RevId: 197027135 | ||
* | Add support to TFLite for dilated convolution. | A. Unique TensorFlower | 2018-04-13 |
| | | | | PiperOrigin-RevId: 192770919 | ||
* | Revert to initializing number of threads when SetNumThreads is called. ↵ | A. Unique TensorFlower | 2018-03-26 |
| | | | | | | | | Requiring it to happen before OpInit() is way too confusing for users. PiperOrigin-RevId: 190499644 | ||
* | Merge changes from github. | Jacques Pienaar | 2018-03-21 |
| | | | | PiperOrigin-RevId: 189945839 | ||
* | Automated g4 rollback of changelist 189231636 | A. Unique TensorFlower | 2018-03-15 |
| | | | | PiperOrigin-RevId: 189258641 | ||
* | Merge changes from github. | Jacques Pienaar | 2018-03-15 |
| | | | | PiperOrigin-RevId: 189231636 | ||
* | Make SetNumThreads apply to the eigen threads. (This creates a dependency on ↵ | A. Unique TensorFlower | 2018-03-09 |
| | | | | | | eigen!) PiperOrigin-RevId: 188504172 | ||
* | Use the multithreaded conv only when threads are available. | A. Unique TensorFlower | 2018-03-09 |
| | | | | PiperOrigin-RevId: 188495357 | ||
* | Adding support for dilated convolution. | A. Unique TensorFlower | 2018-02-21 |
| | | | | PiperOrigin-RevId: 186496353 | ||
* | TFLite Conv2D: Create temporary tensors in Prepare phase. | Yu-Cheng Ling | 2018-02-20 |
| | | | | PiperOrigin-RevId: 186402268 | ||
* | Automated g4 rollback of changelist 186053793 | Yu-Cheng Ling | 2018-02-16 |
| | | | | PiperOrigin-RevId: 186075274 | ||
* | TFLite Conv2D: Create temporary tensors in Prepare phase. | Yu-Cheng Ling | 2018-02-16 |
| | | | | PiperOrigin-RevId: 186053793 | ||
* | Makefile flag to use Apple Accelerate for Conv on iOS. | Yu-Cheng Ling | 2018-02-07 |
| | | | | PiperOrigin-RevId: 184888096 | ||
* | Proper reallocation of dynamic tensors. | A. Unique TensorFlower | 2018-02-05 |
| | | | | PiperOrigin-RevId: 184550199 | ||
* | TFLite: Conv CBLAS kernel | Yu-Cheng Ling | 2018-02-02 |
| | | | | PiperOrigin-RevId: 184328848 | ||
* | Test all TFLite kernel implementations (reference/optimized/...) | Yu-Cheng Ling | 2018-02-01 |
| | | | | PiperOrigin-RevId: 184077940 | ||
* | Fix TFLite Conv reference implementation | Yu-Cheng Ling | 2018-01-22 |
| | | | | PiperOrigin-RevId: 182802058 | ||
* | Internal Change. | Andrew Selle | 2017-11-10 |
PiperOrigin-RevId: 175307445 |