| aeb4531cff7ab757e73f5a910510ebaa32550e59 |
|
23-Feb-2018 |
Miao Wang <miaowang@google.com> |
Merge remote-tracking branch 'aosp/upstream-master' into mymerge
Test: mm
Test: NeuralNetworksTests pass
Change-Id: I347f73fa231f7950eff6a9f7f97730420637b9ec
|
| bd9224f5a066be8ec591bb2ac79c8bd87a9a395b |
|
17-Feb-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Modify reference quantized LSTM implementation so that it only needs one instantiation of fixed-point Tanh, for 3 integer bits, regardless of the value of StateIntegerBits
PiperOrigin-RevId: 186075161
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 0f063cab519f582e2c84560fe1bd82f58f5095c8 |
|
16-Feb-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Optimized quantized LSTM cell runtime NEON implementation.
Notice: unlike many NEON paths that we have in this optimized_ops.h file,
which are enabled also on x86 by means of arm_neon_sse.h (#ifdef USE_NEON),
this one is only enabled on real NEON (#ifdef GEMMLOWP_NEON). The reason
for that is that gemmlowp's FixedPoint class is templatized in the
underlying raw integer/register type, e.g. here int16x8_t, and on SSE
there is only a single __m128i type for all integer types (both int16x8_t
and int32x4_t), making it non-trivial to support this on SSE without
contriving this code on NEON.
PiperOrigin-RevId: 186031054
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 41a8560e0c2fa3b8fa73622a15da53f5e1b8b6c8 |
|
15-Feb-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Implement Split
PiperOrigin-RevId: 185904437
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 90159c53b83527bd088452769ea4e1b98667860c |
|
14-Feb-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Supports op exp (tf.exp) in Toco and Tensorflow Lite.
PiperOrigin-RevId: 185747281
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 39944db497317824ccfd6e9d04d3ea1a6b8c7851 |
|
07-Feb-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Adding support for standalone Tanh operator.
PiperOrigin-RevId: 184845130
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 0255d9a9ca66d233ca4befcceff5d6e70b16d0e0 |
|
07-Feb-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Support for quantized LstmCell, with initial reference runtime code.
The current 'optimized' code is just a copy of the reference code,
a true optimized implementation will follow separately.
PiperOrigin-RevId: 184830223
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 95ed84d5eaa4acccb2f28bda5e5e459921568c3b |
|
05-Feb-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Internal Change
PiperOrigin-RevId: 184532417
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 3cf771cc92402b1b52ea3d1f4b29046e4a7927b2 |
|
05-Feb-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Support for quantized LSTM models.
PiperOrigin-RevId: 184476753
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 1fcb66651d386e1430b3e83a3ce3d971256ffe70 |
|
29-Jan-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Support shrink_axis_mask argument of StridedSlice Op for TfLite.
PiperOrigin-RevId: 183709796
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| b5f9ec4464e0e553c18f61b258a45367030b6b63 |
|
26-Jan-2018 |
Miao Wang <miaowang@google.com> |
Merge remote-tracking branch 'aosp/upstream-master' into mymerge
Test: mm
Change-Id: I5ed4525765fed4d1da4effc8474444eecc22e452
|
| 7b44779ff2031e73ea46ecc7cf2a73405d22a57a |
|
25-Jan-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Prepare internal kernels for BroadcastMul and BroadcastAdd
PiperOrigin-RevId: 183160907
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| f8347ceebbad0e06552633fcdf8e63f52246ba62 |
|
24-Jan-2018 |
Sanjoy Das <sanjoy@google.com> |
Remove THIRD_PARTY_ from #include guards
They don't make sense in the open source repository.
PiperOrigin-RevId: 183140889
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 567e3d15b6e2240c85092d22a8e512b181f5ef6f |
|
23-Jan-2018 |
Miao Wang <miaowang@google.com> |
Merge remote-tracking branch 'aosp/upstream-master' into update_tflite
Test: mm
Change-Id: I58eaf3dd1178e750fef213a8a25ebe1384a4b358
|
| 6dfe00ca4114371ed47c93810219736e3deda2d2 |
|
24-Jan-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Support StridedSlice in TFLite for 1D-4D tensors.
PiperOrigin-RevId: 183020501
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 8afc11ed069ab53760417d8bea7f6d3ca6845597 |
|
17-Jan-2018 |
Andrew Selle <aselle@google.com> |
Create support for div and sub in tensorflow lite.
PiperOrigin-RevId: 182226574
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| a663e5fda16c0c739eb11279422549d499bad597 |
|
11-Jan-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Adds Mean op to Tensorflow Lite.
PiperOrigin-RevId: 181658399
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| fe09efbc7b58be26ac8037bb777053302f5130c2 |
|
08-Jan-2018 |
Nupur Garg <nupurgarg@google.com> |
Support Transpose in TFLite.
The internal implementation supports 1D-4D tensors.
PiperOrigin-RevId: 181221674
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 7f29bfe1903c42bf1caa0806feee5cd579b96311 |
|
02-Jan-2018 |
Nupur Garg <nupurgarg@google.com> |
Support SpaceToBatchND in TFLite.
The internal implementation only supports 4D tensors for now.
The dimension has to be 1 batch + 2 spatial + 1 other.
The most common format within this restriction is NHWC.
PiperOrigin-RevId: 180592870
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 02a4bc4ed24aca9eaedd3a24197244a32b3181f7 |
|
02-Jan-2018 |
Andrew Selle <aselle@google.com> |
Implement reference transpose kernel.
PiperOrigin-RevId: 180571064
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 9393d604ebd63664a27d28617f95fa5f60495270 |
|
15-Dec-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
internal change
PiperOrigin-RevId: 179217499
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 43547b6170c00c2c41478663f350301aa00985ad |
|
07-Dec-2017 |
Miao Wang <miaowang@google.com> |
Create blueprint build file for tflite
- Due to the missing third_party folder, the import path of Eigen
headers was also modified.
- Also provide a static version built with SDK for apps to use.
Test: mm
Change-Id: If0b797a1824282f9527830a0a7f4cf75ae7fa532
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|
| 0b15439f8f0f2d4755587f4096c3ea04cb199d23 |
|
10-Nov-2017 |
Andrew Selle <aselle@google.com> |
Internal Change.
PiperOrigin-RevId: 175307445
/external/tensorflow/tensorflow/contrib/lite/kernels/internal/reference/reference_ops.h
|