About updating to tensorflow-icg based on TF version >= 2.0#6

Open

nbwuzhe wants to merge 10000 commits into

VLOGroup:masterfrom

This pull request is big! We're only showing the most recent 250 commits

Commits on Aug 22, 2019

Commits on Aug 23, 2019

* Add logic to compare computation layout.
tensorflower-gardener
committed
Fix an issue for ExponentialMovingAverage with multiple replicas in distribution strategy.

Yuefeng Zhou
authored and
tensorflower-gardener
committed
make the move semantics more clear. avoid std::move string which is undesired.
tensorflower-gardener
committed
Export the utils functions from C++ to Python with pybind11 instead of swig. This is part of a larger effort to deprecate swig and eventually with modularization break pywrap_tensorflow into smalle…

Amit Patankar
authored and
tensorflower-gardener
committed
Handle IdentityN in TFLiteConverter

miaout17
authored and
tensorflower-gardener
committed
expose profiler locker for other kind of profiler.
tensorflower-gardener
committed
Adds ability to specify an NNAPI execution preference to TF Lite benchmark tool
tensorflower-gardener
committed
Handle If and While ops in TFLite importer
tensorflower-gardener
committed
Automated rollback of commit fd4cfa59c6ba9ead9c6feac0e21f086df27521d2
tensorflower-gardener
committed
[spirv] Add support for extension (de)serialization

antiagainst
authored and
tensorflower-gardener
committed
Add a note to ignore dlerror for CPU-only pip package users when dynamically opening cudart fails.

yifeif
authored and
tensorflower-gardener
committed
Choose best greedy strategy for memory allocation of GL buffers and
tensorflower-gardener
committed
Automated g4 rollback of commit a377b1b43e5a9433a55b86d4af4b48b396177774.

bmzhao
authored and
tensorflower-gardener
committed
Respect dimension id with prediction ops.

tanzhenyu
authored and
tensorflower-gardener
committed
Legalize IdentityN away when transforming to TFL

jpienaar
authored and
tensorflower-gardener
committed
Automated rollback of commit 0f65838cb9787c47f41ede1884e72a0144ad2fe0

aaudiber
authored and
tensorflower-gardener
committed
Add TfLite flex delegate with support for TF ops

jdduke
authored and
tensorflower-gardener
committed
Add support for quantized tensors in TFLite flatbuffer importer
tensorflower-gardener
committed
[Fix] fix the logic when comparing the runtime version strings.

haozha111
authored and
tensorflower-gardener
committed
Fix shader computation precision in convolution and depthwise convolution.
tensorflower-gardener
committed
Define `__slots__` on `_ObjectIdentityWrapper` to speed up wrapper creation.

mrry
authored and
tensorflower-gardener
committed
Fix TensorFlow Lite for Microcontrollers speech training notebook

Daniel Situnayake
authored and
tensorflower-gardener
committed
Automated rollback of commit e32d1900e528a985d90226ebb37939cf280c9e6c

annarev
authored and
tensorflower-gardener
committed
optimize fixed point lstm by precomputing input_zp * weight.

renjie-liu
authored and
tensorflower-gardener
committed
Update Eigen to https://bitbucket.org/eigen/eigen/commits/ce9508807f3fa0ee1244ea80f3f0e695ab14bef2
tensorflower-gardener
committed
Add iterator support to ElementsAttr and SparseElementsAttr.

River707
authored and
tensorflower-gardener
committed
Add I32ElementsAttr to OpBase

karimnosseir
authored and
tensorflower-gardener
committed
Remove unused function.

bmzhao
authored and
tensorflower-gardener
committed
Automated rollback of commit d9e313d10790ae17d0eabbf6e63463510388e182

annarev
authored and
tensorflower-gardener
committed
use I32ElementsAttr from OpBase.

karimnosseir
authored and
tensorflower-gardener
committed
Switch the OSS build to C++14 by default.

jpienaar
authored and
tensorflower-gardener
committed
Update core/distributed_runtime to use tstring.

gharibian
authored and
tensorflower-gardener
committed
Port ADD operation to Micro

YoungSeok Yoon
authored and
tensorflower-gardener
committed
Remove references to contrib in python/ops/variables.py

gunan
authored and
tensorflower-gardener
committed
Correctly calculate the number of elements for CPU memory.
tensorflower-gardener
committed
Update GraphDef version to 136.
tensorflower-gardener
committed
compat: Update forward compatibility horizon to 2019-08-23
tensorflower-gardener
committed
Add initial flow for MlirCompiler::RunBackend.
tensorflower-gardener
committed
Use typedef struct
mattn
committed
Improve shape inference checking for convolutions with batch_group_count > 1.

akuegel
authored and
tensorflower-gardener
committed
Rename HloDialectEmitter to LhloDialectEmitter to better reflect what it does.
tensorflower-gardener
committed
[XLA] Update dependencies on jit.experimental_jit_scope

d0k
authored and
tensorflower-gardener
committed
Remove references to contrib in boosted_trees.

gunan
authored and
tensorflower-gardener
committed
[tf.data] Split Grappler-using utilities out from dataset_utils.cc.

mrry
authored and
tensorflower-gardener
committed
Updated compiler/xrt to use tstring.

gharibian
authored and
tensorflower-gardener
committed
Support Emscripten (ie typically Wasm).

bjacob
authored and
tensorflower-gardener
committed
Remove interface copy. Same interface is defined in gpu/spi.h
tensorflower-gardener
committed
Remove reference to contrib.lookup_ops in iterator_cluster_test.

gunan
authored and
tensorflower-gardener
committed
Dont't swallow exceptions from custom gradients

allenlavoie
authored and
tensorflower-gardener
committed
Contrib is being removed. Remove references to contrib in tensorflow.bzl

gunan
authored and
tensorflower-gardener
committed
Remove unused visisbility rule in toco/python/BUILD

gunan
authored and
tensorflower-gardener
committed
fix OSS issue with cupti tracer
tensorflower-gardener
committed
Cache the name property in `RefVariable`.

mrry
authored and
tensorflower-gardener
committed
restrict profiler_util's visibility
tensorflower-gardener
committed
give profiler interface/session options to control behavior.
tensorflower-gardener
committed
Avoid getting quantization parameters from constant tensors if used by the same scale ops

liufengdb
authored and
tensorflower-gardener
committed
Remove references to contrib in kernel_tests/BUILD.

gunan
authored and
tensorflower-gardener
committed
Fix BufferAllocOp builder.

nicolasvasilache
authored and
tensorflower-gardener
committed
[spirv] NFC: move arithmetic and logical ops to separate files

antiagainst
authored and
tensorflower-gardener
committed
Fix an assertion.

bjacob
authored and
tensorflower-gardener
committed
NFC: Update in-code documentation. Make the two grammar definitions of static-dimension-list consistent.

alinas
authored and
tensorflower-gardener
committed
Introduce the ability for "isolated from above" ops to introduce shadowing
tensorflower-gardener
committed
Merge pull request #31845 from mattn:fix-build-gpu
tensorflower-gardener
committed
Remove references to contrib in toco_flags.proto

gunan
authored and
tensorflower-gardener
committed
Add op to name mapper for exporter

jpienaar
authored and
tensorflower-gardener
committed
[spirv] NFC: move SPIR-V control flow ops to a separate file

antiagainst
authored and
tensorflower-gardener
committed
Update Linalg slice and subview documentation - NFC

nicolasvasilache
authored and
tensorflower-gardener
committed
Add lowering of linalg.copy to an external C++ library and a test.

nicolasvasilache
authored and
tensorflower-gardener
committed
Avoid copying a graph when creating new OpenCL InferenceBuilder.
tensorflower-gardener
committed
Automated rollback of commit 3836459e6112b8fb762fc27b28a80b0d9fd37421
tensorflower-gardener
committed
Add support to only quantize specified operators in the quantization tool.

Suharsh Sivakumar
authored and
tensorflower-gardener
committed
NFC: Add a note to 'applyPatternsGreedily' that it also performs folding/dce.

River707
authored and
tensorflower-gardener
committed
Delete the examples/how_tos directory.

MarkDaoust
authored and
tensorflower-gardener
committed
Add method to only quantize a single layer.

Suharsh Sivakumar
authored and
tensorflower-gardener
committed
Automated rollback of commit 11d96988db9b9303ca47a61cff3ee5c07478e704

aaroey
authored and
tensorflower-gardener
committed
Add L2Norm transformation for TFLite MLIR converter

karimnosseir
authored and
tensorflower-gardener
committed
Use GREEDY_BEST memory strategy in Metal delegate.
tensorflower-gardener
committed
Recursively find all (possibly nested) SymbolRefAttr in TPU program rather than just first-layer ones.

gmagogsfm
authored and
tensorflower-gardener
committed
Disable flaky training_generator_test

gunan
authored and
tensorflower-gardener
committed
Forwardprop: Add utilities for temporarily pushing forward accumulator state

allenlavoie
authored and
tensorflower-gardener
committed
Preserve shape information when passing SparseTensors to dataset functions

aaudiber
authored and
tensorflower-gardener
committed
Remove unused dependency on contrib/util

gunan
authored and
tensorflower-gardener
committed
Automated rollback of commit 1171258036e73c911d0487a3c2db8272fd9dc6be

annarev
authored and
tensorflower-gardener
committed
Added tests and workaround for unsigned integers
tensorflower-gardener
committed
Add dataset ops srcs to :android_all_ops

sdierauf
authored and
tensorflower-gardener
committed
Add an explanation about using FlatbufferModel::GetMinimumRuntime().

haozha111
authored and
tensorflower-gardener
committed
Make NodeNameMapping::Uniquify more efficient.

Xiao Yu
authored and
tensorflower-gardener
committed
Change string to std::string

jpienaar
authored and
tensorflower-gardener
committed
Change behavior of PriorityReadyManager to allow non-unique priorities, and fall back to FirstReadyManager when there is a tie.

wonglkd
authored and
tensorflower-gardener
committed
Adds NMS reference kernel for TFLite to support NonMaxSuppressionV4 & NonMaxSuppressionV5

Sachin Joglekar
authored and
tensorflower-gardener
committed
Convolution 1x1 FP32 specialization for PowerVR.
tensorflower-gardener
committed
Add a linalg.transpose op

nicolasvasilache
authored and
tensorflower-gardener
committed
Rename Greedy to Greedy in order
tensorflower-gardener
committed
Add ShapeN to MLIR Op Definition Spec

andyly
authored and
tensorflower-gardener
committed
Change the title to RC

yashk2810
authored and
tensorflower-gardener
committed
This change is the first step of allowing EagerService executing ops asynchronously:

Xiao Yu
authored and
tensorflower-gardener
committed
Add int8 support to fully_connected kernel on micro.

Nat Jeffries
authored and
tensorflower-gardener
committed
Add custom serializable name support to Model/Sequential.
tensorflower-gardener
committed
Retry "PR #31106: speedup reduce op grads when keep_dims=True"

alextp
authored and
tensorflower-gardener
committed
Document concurrency logic in ScatterNdUpdateOp.

jessanhq
authored and
tensorflower-gardener
committed
Changed return type of ByteCount() to int64_t for convergence between internal/external protobuf.
tensorflower-gardener
committed
Automated rollback of commit dcb8aa7d83cad848473b99f66146b48e8eb26d50
tensorflower-gardener
committed
Exclude python so that the generator doesn't recurse into it.

yashk2810
authored and
tensorflower-gardener
committed
Updates InterpreterOptions, TensorShape, and TensorDataType to be nested types.
tensorflower-gardener
committed
[MLIR/XLA] A round of renames for HLO and LHLO dialects.

cheshire
authored and
tensorflower-gardener
committed

Commits on Aug 24, 2019

Exclude more stuff

yashk2810
authored and
tensorflower-gardener
committed
Lower linalg.transpose to LLVM dialect

nicolasvasilache
authored and
tensorflower-gardener
committed
Automated rollback of commit 91eeb068b1f2709784aba37ec32bb867351bc2be

gunan
authored and
tensorflower-gardener
committed
Lower linalg.copy to LLVM dialect in the presence of transposes.

nicolasvasilache
authored and
tensorflower-gardener
committed
Only truncate ".brain" suffix in eager mode.

rxsang
authored and
tensorflower-gardener
committed
Clarify tests and remove tests for unsupported types.

Nat Jeffries
authored and
tensorflower-gardener
committed
Merge pull request #30202 from freddan80:pr_0626
tensorflower-gardener
committed
NFC: Add doc for id-punct

alinas
authored and
tensorflower-gardener
committed
Move numbers library from lib/strings to platform

gunan
authored and
tensorflower-gardener
committed
Fixed segmentaion fault in ConvPowerVR.
tensorflower-gardener
committed
Update GraphDef version to 137.
tensorflower-gardener
committed
compat: Update forward compatibility horizon to 2019-08-24
tensorflower-gardener
committed
Merge pull request #31924 from mattn:fix-typedef-struct
tensorflower-gardener
committed

Commits on Aug 26, 2019

Updates to Fully Connected GL kernel
tensorflower-gardener
committed
add another scratch tensor for fully quantized lstm, this will be used for store intermediates.

renjie-liu
authored and
tensorflower-gardener
committed
Add neon optimized methods for fixed quantized lstm except MatrixBatchVectoryMultiplyAccumulate method

abattery
authored and
tensorflower-gardener
committed
Upgrade windows RBE toolchain to msvc2017 and python 37.

gunan
authored and
tensorflower-gardener
committed
Add kernel test and tensor util test for LSTM.

jianlijianli
authored and
tensorflower-gardener
committed
Optimize MatrixBatchVectorMultiplyAccumulate for fixed point lstm.

renjie-liu
authored and
tensorflower-gardener
committed
:

pifon2a
authored and
tensorflower-gardener
committed
Fix neon types in neon_tensor_utils

renjie-liu
authored and
tensorflower-gardener
committed
MLIR convert composite function pass, this pass is used to convert to tflite fused ops.

renjie-liu
authored and
tensorflower-gardener
committed
compat: Update forward compatibility horizon to 2019-08-26
tensorflower-gardener
committed
Update GraphDef version to 139.
tensorflower-gardener
committed
Merge pull request #30775 from AyanmoI:amoitra/cudnn_backprop_input_grouped_conv
tensorflower-gardener
committed
Format BUILD in third_party/tensorflow/compiler/mlir/xla.

pifon2a
authored and
tensorflower-gardener
committed
Updated tools/ to use tstring.

gharibian
authored and
tensorflower-gardener
committed
Updated contrib/bigtable/ to use tstring.

gharibian
authored and
tensorflower-gardener
committed
Ruy: Ensure that a couple of classes defined (empty) for non-x86 and non-ARM.

jalexstark
authored and
tensorflower-gardener
committed
Automated rollback of commit b0c2468b4b84ce3ae443390ea5d84cb6e950aee7

jalexstark
authored and
tensorflower-gardener
committed
Use GetTensorData instead of accessing raw data directly.

renjie-liu
authored and
tensorflower-gardener
committed
Merge pull request #28631 from fsx950223:callback
tensorflower-gardener
committed
Updated boosted_trees/training_ops.cc data/dataset_ops.cc to use tstring.

gharibian
authored and
tensorflower-gardener
committed
Updated python/ to use tstring.

gharibian
authored and
tensorflower-gardener
committed
Support folding of ops with inner ops in GreedyPatternRewriteDriver.

andyly
authored and
tensorflower-gardener
committed
Rename xla::mlir to xla::mlir_gpu.

pifon2a
authored and
tensorflower-gardener
committed
Merge pull request #31961 from pwais:patch-1
tensorflower-gardener
committed
Use `Graph._create_op_internal()` instead of `Graph.create_op()`.

mrry
authored and
tensorflower-gardener
committed
Optimize TF_Output construction when building `tf.Tensor` objects.

mrry
authored and
tensorflower-gardener
committed
Changes to merge into most recent tensorflow master
mckib2
committed
Update training_ops.cc
mckib2
authored

Commits on Aug 27, 2019