Jared Duke 23d482eaa2 Add flag for using optimized TFLite CPU kernels on Android ... Add an experimental flag which allows opting in to a set of highly optimized floating point kernels provided via the XNNPACK delegate. This is offered as a preview, with the plan to enable these kernels by default in a future release. The flag can be enabled via: Interpreter.Options options = new Interpreter.Options().setUseXNNPACK(true); See tensorflow/lite/delegates/xnnpack/README.md for more details about these kernels and the associated delegate functionality. PiperOrigin-RevId: 316909226 Change-Id: Ib60cf259225b8a48a9830ccbb24ec10534b038ce		2020-06-17 10:10:58 -07:00
..
main	Add flag for using optimized TFLite CPU kernels on Android	2020-06-17 10:10:58 -07:00
test	Add flag for using optimized TFLite CPU kernels on Android	2020-06-17 10:10:58 -07:00
testdata	Improve image testing for TFLite Java tests	2020-05-13 13:16:18 -07:00
testhelper/java/org/tensorflow/lite	Apply 'buildozer fix moveLicensesAndDistribs movePackageToTop' to all BUILD files.	2019-05-24 04:53:01 -07:00