tensorflow / tensorflow

74 Pull requests merged by 16 people

• Fix wrong instruction of TFLite iOS profiling tool

  #51079 merged Aug 2, 2021

• Prevent heap OOB read in TFLite's `gather.cc`.

  #51018 merged Jul 30, 2021

• Prevent division by 0 in LSH projection.

  #51017 merged Jul 30, 2021

• Prevent nullptr dereference in MLIR TFLite dialect/optimizer.

  #51016 merged Jul 30, 2021

• Prevent heap OOB read in TFLite's `gather_nd.cc`.

  #51014 merged Jul 30, 2021

• Prevent an OOB read in `expand_dims.cc`

  #51013 merged Jul 30, 2021

• Disallow negative ngram_widths values in tf.raw_ops.StringNGrams

  #51022 merged Jul 30, 2021

• Disallow negative ngram_widths values in tf.raw_ops.StringNGrams

  #51021 merged Jul 30, 2021

• Disallow negative ngram_widths values in tf.raw_ops.StringNGrams

  #51020 merged Jul 30, 2021

• Disallow negative ngram_widths values in tf.raw_ops.StringNGrams

  #51019 merged Jul 30, 2021

• Ensure non-empty input_splits in tf.raw_ops.UnicodeEncode

  #51012 merged Jul 30, 2021

• Ensure non-empty input_splits in tf.raw_ops.UnicodeEncode

  #51011 merged Jul 30, 2021

• Ensure non-empty input_splits in tf.raw_ops.UnicodeEncode

  #51010 merged Jul 30, 2021

• Ensure non-empty input_splits in tf.raw_ops.UnicodeEncode

  #51009 merged Jul 30, 2021

• Secure tf.raw_ops.QuantizeV2

  #51008 merged Jul 30, 2021

• Secure tf.raw_ops.QuantizeV2

  #51007 merged Jul 30, 2021

• Secure tf.raw_ops.QuantizeV2

  #51006 merged Jul 30, 2021

• Secure tf.raw_ops.QuantizeV2

  #51005 merged Jul 30, 2021

• [CherryPick2.6]Fix bug that could cause map_fn to produce incorrect results (rather than an error) when mapping over a ragged tensor with an inappropriate fn_output_signature. (Note: there are cases where the default value for fn_output_signature is not appropriate, so the user needs to explicitly specify the correct output signature.)

  #51050 merged Jul 30, 2021

• [Cherrypick2.6] In tf.raw_ops.BoostedTreesSparseCalculateBestFeatureSplit, limit stat_dim in stats_summary_indices to under stats_dims in stats_summary_shape

  #51004 merged Jul 30, 2021

• [Cherrypick2.5] In tf.raw_ops.BoostedTreesSparseCalculateBestFeatureSplit, limit stat_dim in stats_summary_indices to under stats_dims in stats_summary_shape

  #51003 merged Jul 30, 2021

• [Cherrypick2.4] In tf.raw_ops.BoostedTreesSparseCalculateBestFeatureSplit, limit stat_dim in stats_summary_indices to under stats_dims in stats_summary_shape

  #51002 merged Jul 30, 2021

• [Cherrypick2.3] In tf.raw_ops.BoostedTreesSparseCalculateBestFeatureSplit, limit stat_dim in stats_summary_indices to under stats_dims in stats_summary_shape

  #51001 merged Jul 30, 2021

• [Cherrypick2.6] Disallow empty node_id_range in tf.raw_ops.BoostedTreesCalculateBestFeatureSplitV2 and tf.raw_ops.BoostedTreesCalculateBestGainsPerFeature

  #50997 merged Jul 30, 2021

• [Cherrypick2.3] Disallow empty node_id_range in tf.raw_ops.BoostedTreesCalculateBestFeatureSplitV2 and tf.raw_ops.BoostedTreesCalculateBestGainsPerFeature

  #50998 merged Jul 30, 2021

• [Cherrypick2.4] Disallow empty node_id_range in tf.raw_ops.BoostedTreesCalculateBestFeatureSplitV2 and tf.raw_ops.BoostedTreesCalculateBestGainsPerFeature

  #50999 merged Jul 30, 2021

• [Cherrypick2.5] Disallow empty node_id_range in tf.raw_ops.BoostedTreesCalculateBestFeatureSplitV2 and tf.raw_ops.BoostedTreesCalculateBestGainsPerFeature

  #51000 merged Jul 30, 2021

• TF-TRT Apply batch size check for FC conversion path of MatMul

  #50947 merged Jul 29, 2021

• [TFLite] Add TABLE operator for LUT-based operators lowering

  #45342 merged Jul 29, 2021

• Update gast mirror URL

  #50882 merged Jul 29, 2021

• use `std::make_tuple` in `cupti_tracer.cc`

  #50987 merged Jul 28, 2021

• [oneDNN] Fix for build failure.

  #50991 merged Jul 28, 2021

• Revert "Update shape_refiner.cc"

  #50996 merged Jul 28, 2021

• Fix segmentation fault in shape inference logic.

  #50985 merged Jul 28, 2021

• Fix segmentation fault in shape inference logic.

  #50984 merged Jul 28, 2021

• Fix segmentation fault in shape inference logic.

  #50983 merged Jul 28, 2021

• Add missing closing backtick in parse_sequence_example docstring

  #50976 merged Jul 28, 2021

• Fix numpy 1.20 deprecation warnings (round two)

  #50930 merged Jul 28, 2021

• [Go] replace deprecated protocol buffers module

  #50312 merged Jul 28, 2021

• Enable '--distinct_host_configuration' for arm builds

  #50893 merged Jul 27, 2021

• 2.6rc2 cherry-pick request: Fix issue where disabling TensorFloat-32 had no effect.

  #50879 merged Jul 27, 2021

• [CherryPick2.6]Fix segfault on string tensors with mismatched dimensions

  #50878 merged Jul 27, 2021

• [CherryPick2.6]Prevent infinite loop in strided slice implementation

  #50868 merged Jul 27, 2021

• [CherryPick2.6]Prevent a division by 0 in average ops.

  #50867 merged Jul 27, 2021

• [CherryPick2.6]Prevent a division by 0 in division ops.

  #50866 merged Jul 27, 2021

• [CHerryPick2.6]Fix a null pointer exception caused by branching on uninitialized data.

  #50865 merged Jul 27, 2021

• [CherryPick2.6]Fix a null pointer exception caused by branching on uninitialized data.

  #50864 merged Jul 27, 2021

• [CherryPick2.6]Fix a null pointer exception caused by branching on uninitialized data.

  #50863 merged Jul 27, 2021

• [CherryPick:2.6]Fix a null pointer exception in SVDF

  #50862 merged Jul 27, 2021

• [CherryPick2.6]Prevent division by 0 in `fully_connected.cc`

  #50861 merged Jul 27, 2021

• [Cherrypick:2.6]Disallow dims input of 0 in tf.raw_ops.UnravelIndex

  #50858 merged Jul 27, 2021

• use `std::make_tuple` instead of non-empty braced-init-lists

  #50929 merged Jul 27, 2021

• [INTEL MKL] Update mkl auto_mixed_precision_lists

  #50931 merged Jul 27, 2021

• Added Python 3.9 support for "Building TensorFlow Lite Standalone Pip"

  #50895 merged Jul 27, 2021

• [CherryPick2.6]Ensure non-empty compressed input in tf.raw_ops.UncompressElement

  #50849 merged Jul 26, 2021

• Ensure non-empty compressed input in tf.raw_ops.UncompressElement

  #50912 merged Jul 26, 2021

• Ensure non-empty compressed input in tf.raw_ops.UncompressElement

  #50911 merged Jul 26, 2021

• Ensure non-empty compressed input in tf.raw_ops.UncompressElement

  #50910 merged Jul 26, 2021

• [CherryPick2.6]Validate num_elements input in tf.raw_ops.TensorListReserve

  #50848 merged Jul 26, 2021

• Validate num_elements input in tf.raw_ops.TensorListReserve

  #50909 merged Jul 26, 2021

• Validate num_elements input in tf.raw_ops.TensorListReserve

  #50908 merged Jul 26, 2021

• Validate num_elements input in tf.raw_ops.TensorListReserve

  #50907 merged Jul 26, 2021

• [CherryPick2.6]Fix null ptr deref in tf.raw_ops.RaggedTensorToTensor

  #50847 merged Jul 26, 2021

• Fix null ptr deref in tf.raw_ops.RaggedTensorToTensor

  #50906 merged Jul 26, 2021

• Fix null ptr deref in tf.raw_ops.RaggedTensorToTensor

  #50905 merged Jul 26, 2021

• Fix null ptr deref in tf.raw_ops.RaggedTensorToTensor

  #50904 merged Jul 26, 2021

• Fix accessing possible nullptr in tensorflow::data::CompressElement and UncompressElement which are used in tf.data.service.

  #50903 merged Jul 26, 2021

• Fix accessing possible nullptr in tensorflow::data::CompressElement and UncompressElement which are used in tf.data.service.

  #50902 merged Jul 26, 2021

• Fix accessing possible nullptr in tensorflow::data::CompressElement and UncompressElement which are used in tf.data.service.

  #50901 merged Jul 26, 2021

• Disallow division by zero FPE in tf.raw_ops.SparseDenseCwiseDiv

  #50898 merged Jul 26, 2021

• Disallow division by zero FPE in tf.raw_ops.SparseDenseCwiseDiv

  #50899 merged Jul 26, 2021

• Disallow division by zero FPE in tf.raw_ops.SparseDenseCwiseDiv

  #50900 merged Jul 26, 2021

• [CherryPick2.6]Disallow division by zero FPE in tf.raw_ops.SparseDenseCwiseDiv

  #50846 merged Jul 26, 2021

• add device_vendor to the GpuDevice in JAX runtime

  #50799 merged Jul 26, 2021

34 Pull requests opened by 16 people

• [MLIR][DISC] add ral logger and registry

  #50954 opened Jul 26, 2021

• Add back "PR #49173: [Crash fix] Fix cudaMallocAsync crashes."

  #50961 opened Jul 26, 2021

• Crash fix cudaMallocAsync usage of TF_CUDA_MALLOC_ASYNC_SUPPORTED_PREALLOC.

  #50962 opened Jul 26, 2021

• [PJRT] add platfrom_version for rocm in pjrt/gpu_device

  #50972 opened Jul 27, 2021

• Make passthrough alias work for resource updates

  #50986 opened Jul 27, 2021

• [determinism] Add unimplemented exception to nearest-neighbor resizing

  #51023 opened Jul 29, 2021

• Revert https://github.com/tensorflow/tensorflow/pull/50292

  #51024 opened Jul 29, 2021

• Add hlo to mhlo logistic translation

  #51028 opened Jul 29, 2021

• Remove AWS files as s3 support is now in modular file system

  #51032 opened Jul 29, 2021

• Remove hdfs as support has been moved to modular file systems.

  #51033 opened Jul 29, 2021

• fix: miss typo codespelling

  #51035 opened Jul 29, 2021

• Update RELEASE.md to add entry for modular file system support (s3 and hdfs)

  #51036 opened Jul 29, 2021

• Fix missing preprocess guards in CUDNN wrapper (Fixes build breakage when CUDNN Frontend not used or CUDNN < 8.1)

  #51039 opened Jul 30, 2021

• [MLIR][DISC] Bufferize mhlo::LogisticOp

  #51044 opened Jul 30, 2021

• Update max version of bazel supported by tensorflow to 4.99.0

  #51045 opened Jul 30, 2021

• remove cast of y_true to y_pred data type in sparse categorical cross…

  #51053 opened Jul 30, 2021

• Fix segfault/heap buffer overflow in `{Experimental,}DatasetToTFRecord` where dataset is numeric.

  #51054 opened Jul 30, 2021

• Add missing validation to `RaggedTensorToSparse`.

  #51055 opened Jul 30, 2021

• [cheerypick2.6]Ensure num_streams >= 0 in tf.raw_ops.BoostedTreesCreateQuantileStreamResource

  #51056 opened Jul 30, 2021

• Add more validation to `RequantizationRangePerChannel`.

  #51057 opened Jul 30, 2021

• Ensure non-empty rt_nested_splits in tf.raw_ops.RaggedTensorToVariant

  #51058 opened Jul 30, 2021

• [cherryPick2.6]

  #51059 opened Jul 30, 2021

• Prevent CHECK-fail/heap OOB in UpperBound and LowerBound

  #51060 opened Jul 30, 2021

• Prevent overflow due to integer conversion to unsigned.

  #51061 opened Jul 30, 2021

• Prevent nullptr deref in validation of indexes in map ops.

  #51062 opened Jul 30, 2021

• Add remaining validation to `sdca_internal.cc`

  #51063 opened Jul 30, 2021

• Prevent a CHECK-fail due to empty tensor input in `map_stage_op.cc`

  #51064 opened Jul 30, 2021

• Prevent a segfault in shape inference due to bad inputs.

  #51065 opened Jul 30, 2021

• Prevent dereferencing of null pointers in TFLite's `add.cc`.

  #51066 opened Jul 30, 2021

• Fix a shape inference issue leading to nullptr deref.

  #51067 opened Jul 30, 2021

• Prevent division by 0 in common shape functions.

  #51068 opened Jul 30, 2021

• [Cherrypick2.6] Update XNNPACK dependency

  #51069 opened Jul 31, 2021

• add missing dependencies for ChloPasses

  #51085 opened Aug 1, 2021

• Updated README.md

  #51087 opened Aug 2, 2021

82 Issues closed by 38 people

• 16x8 quantization mode & TFLite Micro

  #51089 closed Aug 2, 2021

• SECURITY.md

  #50922 closed Aug 2, 2021

• Incomplete Installation of Tensorflow FreeBSD

  #50766 closed Aug 1, 2021

• NameError: name 'interpreter_wrapper' is not defined

  #50158 closed Aug 1, 2021

• dnn implementation Error

  #51081 closed Aug 1, 2021

• Regarding Scroll for the nearest points

  #50808 closed Jul 30, 2021

• Serialize tf.keras Preprocessing layers

  #50793 closed Jul 30, 2021

• layer order in functional API graph models

  #50306 closed Jul 30, 2021

• [_Derived_]RecvAsync is cancelled while running tensorflow RNN tutorial code on GPU

  #50296 closed Jul 30, 2021

• FasterRCNN_resnet101 .tflite conversion

  #37401 closed Jul 30, 2021

• Dense variable constraints are not allowed with sparse gradients

  #33755 closed Jul 30, 2021

• `LLVM ERROR: Cannot select` issue when compiling LLVM module on s390x architecture

  #44362 closed Jul 30, 2021

• Release candidate 2.6.0rc1 is not built with HDFS support anymore?

  #51046 closed Jul 30, 2021

• Inception_v4 can not node in the pb file

  #51047 closed Jul 30, 2021

• Cannot save custom model

  #50247 closed Jul 30, 2021

• Android model personalization crash with keras model

  #37546 closed Jul 30, 2021

• AttributeError: 'tensorflow.python.framework.ops.EagerTensor' object has no attribute 'decode'

  #50525 closed Jul 30, 2021

• Training becomes increasingly slower when creating models in a loop

  #47971 closed Jul 30, 2021

• Worse predict result from keras h5 model convert to tensorflow frozen pb model

  #51027 closed Jul 30, 2021

• Build TensorFlow Lite for Android and reduce binary size to use with C API

  #50946 closed Jul 30, 2021

• In distributed training, different workers get different results from the information

  #50746 closed Jul 29, 2021

• google.protobuf.message.DecodeError With Multi-GPU and Mirrored Strategy on 2.2rc3 (and 2.2rc4)

  #39082 closed Jul 29, 2021

• NotImplementedError: Cannot convert a symbolic Tensor (lstm_20/strided_slice:0) to a numpy array...

  #50784 closed Jul 29, 2021

• Problems on TensorFlow-ROCm installation

  #50944 closed Jul 29, 2021

• XLA Optimizer Mixed Precision Support

  #50981 closed Jul 29, 2021

• Tflite: issue with dilated convolution in a dynamic shape CNN

  #50448 closed Jul 29, 2021

• gast_archive URL error

  #50777 closed Jul 29, 2021

• Help Me 😅

  #50990 closed Jul 29, 2021

• Failure to load model files.

  #49162 closed Jul 29, 2021

• At Chrome run Tensorflow.js.

  #50968 closed Jul 29, 2021

• fake quant in tflite ignores the num_bits that is passed

  #50993 closed Jul 29, 2021

• KerasTensors do not support tf.custom_gradient()

  #50741 closed Jul 28, 2021

• TF2 GPUs on Kubernetes

  #45494 closed Jul 28, 2021

• Converted models gives wrong outputs

  #46495 closed Jul 28, 2021

• Accept h5py.Group in tf.keras save/load functions

  #25181 closed Jul 28, 2021

• Tensorflow GPU installation don't want to run this code

  #50614 closed Jul 28, 2021

• Return gradient 0 in mixed_precison setting

  #47884 closed Jul 28, 2021

• ValueError: Could not find matching function to call loaded from the SavedModel

  #50928 closed Jul 28, 2021

• Tensorflow 2.3 GPU build failed | NCCL no such package | RHEL 7

  #49377 closed Jul 28, 2021

• Unknown: input/output error

  #50824 closed Jul 28, 2021

• tf.compat.v1.profiler.profile produces wrong flops on Conv2D

  #50802 closed Jul 28, 2021

• apply_gradients() with Adam Optimizer gives error

  #50769 closed Jul 28, 2021

• tf.keras.layers.TimeDistributed build method passes shapes as tuples instead of TensorShapes.

  #50752 closed Jul 28, 2021

• Custom metric fails with "ValueError: Metric should be a callable"

  #50749 closed Jul 28, 2021

• Building TF.Lite benchmark tool with TF.text ops support

  #50924 closed Jul 28, 2021

• bug

  #50963 closed Jul 28, 2021

• I'm working in colab and changed `tf.gfile.GFile` to `tf.io.gfile.GFile` but still get this error

  #50809 closed Jul 28, 2021

• Tensorflow does not use the GPU during training with eager execution, despite manual activation

  #45546 closed Jul 28, 2021

• ModuleNotFounderror:no module named tensorflor_core.estimator

  #50763 closed Jul 28, 2021

• Missing libtensorflow builds for 2.5.0-rc0 and 2.5.0-rc1

  #48550 closed Jul 28, 2021

• Keras.io XRay TPU example fails on 2.6.0-rc1

  #50971 closed Jul 28, 2021

• Build TFLite Micro for RISC-V

  #47622 closed Jul 28, 2021

• Unable to convert mT5 model to tflite (tensorflow.GraphDef exceeds maximum protobuf size of 2GB)

  #47326 closed Jul 28, 2021

• FunctionalPreprocessingStage incompatible with some keras preprocessing layers

  #50979 closed Jul 28, 2021

• Could not find org.tensorflow:tensorflow-lite-support:0.1.0-rc1

  #50896 closed Jul 28, 2021

• ```tf.keras.callbacks.EarlyStopping``` always stops at the first epoch

  #49227 closed Jul 28, 2021

• Documentation: old installation requirements

  #50973 closed Jul 28, 2021

• SparseCategoricalCrossentropy and Mixed Precision Training

  #50964 closed Jul 28, 2021

• Increasing physical memory usage implemented with tensorflow.

  #50733 closed Jul 28, 2021

• Bug in custom layer input tensor

  #50377 closed Jul 28, 2021

• Micro kernel elementwise math ops issues (abs, sin, cos, log, sqrt, tanh)

  #39020 closed Jul 28, 2021

• ERROR: Did not get operators, tensors, or buffers in subgraph 0.

  #38298 closed Jul 28, 2021

• TFLite MaxPool2D GPU op seems to modify source tensor

  #38732 closed Jul 28, 2021

• RuntimeError: Attempting to capture an EagerTensor without building a function.

  #50978 closed Jul 28, 2021

• XLA compilation bug in TPU involving indirection through index tensors

  #48268 closed Jul 27, 2021

• XNNPACK enable cross-compilation failure using Bazel on r2.6 armhf or v2.6.0-rc1 armhf TensorFlow Lite pip

  #50920 closed Jul 27, 2021

• Failed to convert object of type <class 'tensorflow.python.framework.sparse_tensor.SparseTensor'> to Tensor

  #50959 closed Jul 27, 2021

• None of the Demo Edge TPU Models Work - “Module Provided has model identifier 'l><b', should be 'TFL3'”

  #50974 closed Jul 27, 2021

• Unable to use Keras data generator with Functional API model

  #50745 closed Jul 27, 2021

• Wrong access modifiers in tf::ResourceHandle

  #50481 closed Jul 27, 2021

• Possibility to define operator version

  #50731 closed Jul 27, 2021

• experimental apis in iOS pod is missing.

  #46852 closed Jul 27, 2021

• [Question] TfLiteTensor is supposed to be deleted/released with TfLiteTensorFree?

  #50886 closed Jul 27, 2021

• api_docs tensor incoherent sentence

  #50452 closed Jul 26, 2021

• My model seems not using my custom designed loss function

  #50711 closed Jul 26, 2021

• Saved model loading : KeyError: '__inference_depthwise_conv2d_layer_call_fn_126'

  #50706 closed Jul 26, 2021

• InvalidArgumentError: No OpKernel was registered to support Op 'QuantizedMatMulWithBiasAndDequantize'

  #50686 closed Jul 26, 2021

• Retrieving the Unreduced Losses

  #50704 closed Jul 26, 2021

• add a new class for realizing Early Stopping

  #33739 closed Jul 26, 2021

• ProtoCompile Error while building tf 2.5 from source

  #50734 closed Jul 26, 2021

• 'tf.MLCConv2D' op is neither a custom op nor a flex op

  #50933 closed Jul 26, 2021

• Predict data from a collection to determine a single object with the most used data.

  #50897 closed Jul 26, 2021

51 Issues opened by 49 people

• Single scalar ground truth for a batch?

  #51090 opened Aug 2, 2021

• RTX30 series: unsupported compute compability for CUDA 9.0

  #51088 opened Aug 2, 2021

• Does TFLite Support conv + LSTM ?

  #51086 opened Aug 2, 2021

• Cannot split windowed data when batching is used in tf.data.Dataset

  #51084 opened Aug 1, 2021

• tf.io.read_file memory leackage

  #51083 opened Aug 1, 2021

• NotImplementedError: Cannot convert a symbolic Tensor (digit_capsule_layer_2/strided_slice:0) to a numpy array. This error may indicate that you're trying to pass a Tensor to a NumPy call, which is not supported

  #51082 opened Aug 1, 2021

• Op type not registered 'NormalizeUTF8' in binary running when reload model

  #51080 opened Aug 1, 2021

• cannot use tf.keras.preprocessing.image.random_channel_shift or apply_channel_shift with tf.data.Dataset functions like image_dataset_from_directory or tensor_slices.

  #51078 opened Jul 31, 2021

• tf.math.confusion_matrix InvalidArgumentError about incompatible shapes

  #51077 opened Jul 31, 2021

• __iter__() is only supported inside of tf.function or when eager execution is enabled

  #51076 opened Jul 31, 2021

• Linux setup guide update for Ubuntu 20.04

  #51075 opened Jul 31, 2021

• tf.spars.sparsee_dense_matmul supports multiplication of batches of data in a sparse tensor with a dense tensor

  #51074 opened Jul 31, 2021

• INFO:tensorflow:Assets written to: C:\Users\TFLab\AppData\Local\Temp\tmpdw25wybi\assets

  #51073 opened Jul 31, 2021

• Can sparse_csr_matrix_ops.sparse_tensor_to_csr_sparse_matrix be used in the graph mode?

  #51072 opened Jul 31, 2021

• Add version to Zenodo

  #51071 opened Jul 31, 2021

• Index out of bounds error

  #51070 opened Jul 31, 2021

• issue with running forward passes using saved model format converted from a checkpoint

  #51052 opened Jul 30, 2021

• Traing is very slow when add a custom layers. I found this tensor ops run on cpu not on gpu,I don't konw why?

  #51051 opened Jul 30, 2021

• using `tf.image.*` module for one channel or more than three channel?

  #51049 opened Jul 30, 2021

• Two python versions install tf in one system one succeeds while the other fails

  #51048 opened Jul 30, 2021

• bug in rsqrt

  #51043 opened Jul 30, 2021

• Does the "p" in "pfor" stand for "print" or something?

  #51041 opened Jul 30, 2021

• Bash Bad address if CUDA on None Standard Path

  #51040 opened Jul 30, 2021

• Autograph set_verbosity doc example shows using integer values for env vars, but only strings can be used

  #51038 opened Jul 29, 2021

• Installation of tflite-model-maker 0.3.2. takes forever

  #51031 opened Jul 29, 2021

• TfLite python delegate support for RB5 robotics platform

  #51030 opened Jul 29, 2021

• Query About tflite interpreter usage

  #51029 opened Jul 29, 2021

• TensorFlow Lite C API full example

  #51026 opened Jul 29, 2021

• Could provide the different platforms compiled release file, as widows lib and in(cpu gpu) and linux .so?

  #51025 opened Jul 29, 2021

• Dependency Tree for tflite-model-maker 0.2.5 on Pypi

  #50994 opened Jul 28, 2021

• reduce_sum in tflite outputs is incorrect

  #50992 opened Jul 28, 2021

• Failed when upgrading grpc to 1.36.0

  #50989 opened Jul 28, 2021

• How to convert model trained with tensorflow object detection api to tflite model that supported by Android GPU

  #50988 opened Jul 28, 2021

• Why is <normalizing flow>.prob() failing for me?

  #50982 opened Jul 27, 2021

• XLA compilation error for TPU

  #50980 opened Jul 27, 2021

• Issue with saving trained model of module TensorFlow Recommenders (tfrs)

  #50977 opened Jul 27, 2021

• failed call to cuDevicePrimaryCtxRetain: CUDA_ERROR_INVALID_DEVICE: invalid device ordinal

  #50975 opened Jul 27, 2021

• Node number 106 (FlexTensorScatterUpdate) failed to prepare.

  #50970 opened Jul 27, 2021

• protobuf error while saving the model

  #50967 opened Jul 27, 2021

• The max version of bazel supported by tensorflow should be increased to 4.1.0 or higher

  #50966 opened Jul 27, 2021

• Tensorflow Lite Model Maker cannot train EfficientDet-Lite using Cloud TPU

  #50965 opened Jul 27, 2021

• "ValueError: Cannot take the length of Shape with unknown rank" or "InvalidArgumentError: Length for attr 'output_shapes' of 0 must be at least minimum 1" using tf.data.Dataset tensors

  #50960 opened Jul 26, 2021

• Code not visible in dark theme in the Tensorflow documentation

  #50958 opened Jul 26, 2021

• resize_tensor_input does not change shape in tflite interpreter for efficientdet models

  #50957 opened Jul 26, 2021

• Reading TFRecords from S3 with a window shift leads to a large performance degradation

  #50956 opened Jul 26, 2021

• Could not find the DLL(s) 'msvcp140.dll or msvcp140_1.dll'

  #50955 opened Jul 26, 2021

• Identifying the type of input

  #50953 opened Jul 26, 2021

• Unexpected warning when using tf.data.Dataset.cache(filename)

  #50952 opened Jul 26, 2021

• smaller model runs slower than a larger one when compiled for edgetpu

  #50951 opened Jul 26, 2021

• cl.exe failed during building tensorflow from source

  #50950 opened Jul 26, 2021

• import tensorflow get a bug

  #50948 opened Jul 26, 2021

234 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• [XLA] Enable pointwise row vectorization for small row.

  #50756 commented on Aug 1, 2021 • 8 new comments

• validation_split for tensorflow.dataset

  #50937 commented on Jul 28, 2021 • 5 new comments

• Bug with optimizer_v2.OptimizerV2.set_weights method

  #50859 commented on Jul 29, 2021 • 5 new comments

• Numpy v1.20+ compatibility

  #47691 commented on Jul 29, 2021 • 5 new comments

• Adding some Boosted tree ops to the 'allowed' list

  #50801 commented on Jul 31, 2021 • 4 new comments

• /libmetal_plugin.dylib: Symbol not found: _TF_AssignUpdateVariable

  #50541 commented on Aug 1, 2021 • 4 new comments

• Tensorflow 2.3.x link error when use delay load

  #50658 commented on Jul 28, 2021 • 3 new comments

• build fails on AArch64, Fedora 33 due to --config=numa, missing sysctl.h

  #50574 commented on Jul 29, 2021 • 3 new comments

• AttributeError: 'NoneType' object has no attribute 'PluginAssets'

  #50921 commented on Aug 2, 2021 • 3 new comments

• Error build: target 'cuda/bin/nvcc' not declared in package

  #50943 commented on Aug 2, 2021 • 3 new comments

• Add use_default_shell_env = True to all ctx.actions.run_shell rules

  #44549 commented on Jul 31, 2021 • 3 new comments

• Make tf.image.resize compatible with XLA compilation

  #46517 commented on Aug 2, 2021 • 3 new comments

• [oneDNN] Improving Graph Rewrite Performance

  #50932 commented on Jul 29, 2021 • 3 new comments

• Dynamic-sized CoreML Delegate

  #50758 commented on Jul 27, 2021 • 2 new comments

• Predict on batch being able to do the prediction with inputs of different first shapes

  #42280 commented on Jul 28, 2021 • 2 new comments

• KeyError when taking derivative to input of Conv2D with tf.function only

  #45063 commented on Jul 28, 2021 • 2 new comments

• 2.5.0 tf_to_kernel is not compatible with msvc /Ob3

  #50935 commented on Jul 28, 2021 • 2 new comments

• TF 2.2.0: Error with model.fit; class_weight is only supported for Models with a single output.

  #40457 commented on Jul 29, 2021 • 2 new comments

• Cannot load SavedModel when model trained with tfa.optimizers.NovoGrad: ValueError: Shapes (3, 8) and () are incompatible

  #45557 commented on Jul 29, 2021 • 2 new comments

• Go installation error

  #43847 commented on Jul 29, 2021 • 2 new comments

• Low GPU usage of RNN layers under MirroredStrategy

  #30715 commented on Jul 29, 2021 • 2 new comments

• TF-TRT Dynamic Shapes Feature Tracker

  #45481 commented on Jul 29, 2021 • 2 new comments

• TFlite fails to run on GPU with - ERROR: TfLiteGpuDelegate Prepare: No shader implementation for reduce_sum

  #45467 commented on Jul 29, 2021 • 2 new comments

• Introduce ability to clear GPU memory in Tensorflow 2

  #48545 commented on Jul 29, 2021 • 2 new comments

• Can't save the LSTM model. (AttributeError: 'CuDNNLSTM' object has no attribute 'unroll')

  #50945 commented on Jul 30, 2021 • 2 new comments

• from_dlpack unable to process arrays with column-major strides

  #42660 commented on Jul 30, 2021 • 2 new comments

• model.fit with dataset generator results in deadlock/hang

  #50637 commented on Jul 30, 2021 • 2 new comments

• GPU Profiling: MemoryProfile do not contain memory events.

  #48146 commented on Jul 31, 2021 • 2 new comments

• tensorflow.data.experimental.load not working inside tensorflow.data.Dataset().flatmap

  #50594 commented on Aug 2, 2021 • 2 new comments

• `case` and `cond` executes all the functions within

  #50843 commented on Aug 2, 2021 • 2 new comments

• [Intel MKL] Fix shape inference for Quantized{Conv2D/DepthwiseConv2D}-like operations

  #46779 commented on Aug 2, 2021 • 2 new comments

• Issue tensorflow#33840 in tf.image.ssim_multiscale

  #48938 commented on Jul 29, 2021 • 2 new comments

• [determinism] Add sparse softmax/xent GPU-determinism

  #50070 commented on Aug 1, 2021 • 2 new comments

• [INTEL MKL] Enable mkl group conv

  #50333 commented on Jul 29, 2021 • 2 new comments

• [WIP][MLIR][DISC] Add device type info to memref's address space in lmhlo dialect

  #50365 commented on Jul 29, 2021 • 2 new comments

• Move code shared between lite and micro to elementwise_portable.h

  #50836 commented on Jul 26, 2021 • 2 new comments

• Tensorflow estimator returns the same answer

  #50872 commented on Jul 26, 2021 • 1 new comment

• Where is convert_all_kernels_in_model in tensorflow 2.5.0?

  #50941 commented on Jul 26, 2021 • 1 new comment

• converting tensorflow model to mlir

  #50890 commented on Jul 26, 2021 • 1 new comment

• "Given shapes (...) are not broadcastable" after conversion from tensorflow to tflite

  #50676 commented on Jul 26, 2021 • 1 new comment

• Could not load dynamic library 'cupti64_110.dll'; dlerror: cupti64_110.dll not found

  #43030 commented on Jul 26, 2021 • 1 new comment

• Discrepancy between expected and tflite quantization parameters

  #50857 commented on Jul 26, 2021 • 1 new comment

• TensorFlow Certificate Network - region - australiasia

  #50812 commented on Jul 26, 2021 • 1 new comment

• Cannot convert a symbolic Keras input/output to a numpy array when using TF Lite converter

  #50719 commented on Jul 26, 2021 • 1 new comment

• InvalidArgumentError: indices[0,2] = 2188 is not in [0, 2027)

  #50845 commented on Jul 26, 2021 • 1 new comment

• tensorflow estimator how to export sub-graph

  #50833 commented on Jul 26, 2021 • 1 new comment

• 'utf-8' codec can't decode byte 0xc1 in position 103: invalid start byte

  #50837 commented on Jul 26, 2021 • 1 new comment

• Internal Error with TF_GPU_ALLOCATOR=cuda_malloc_async

  #50669 commented on Jul 26, 2021 • 1 new comment

• AttributeError: 'ReLU' object has no attribute '_saved_model_inputs_spec'

  #50702 commented on Jul 26, 2021 • 1 new comment

• Tensorflow shuffle leak memory

  #50406 commented on Jul 26, 2021 • 1 new comment

• WARNING:absl:Found untraced functions such as lstm_cell_2_layer_call_fn, lstm_cell_2_layer_call_and_return_conditional_losses, lstm_cell_2_layer_call_fn, lstm_cell_2_layer_call_and_return_conditional_losses, lstm_cell_2_layer_call_and_return_conditional_losses while saving (showing 5 of 5). These functions will not be directly callable after loading.

  #47554 commented on Jul 26, 2021 • 1 new comment

• Build TensorFlowLite C API with GPU delegate with CMake

  #50940 commented on Jul 27, 2021 • 1 new comment

• TypeError: Can't convert Operation 'MutableHashTable' to Tensor

  #24439 commented on Jul 27, 2021 • 1 new comment

• AssertionError occurred when initializing NLClassifier with TFlite model

  #50764 commented on Jul 27, 2021 • 1 new comment

• loading a keras model on raspberry pi

  #50915 commented on Jul 27, 2021 • 1 new comment

• BigBiGAN Loss Function

  #50674 commented on Jul 27, 2021 • 1 new comment

• tf.image.ssim_multiscale does not work in tf-2.0.0

  #33840 commented on Jul 27, 2021 • 1 new comment

• A custom RNN cell does not support a high-order state_size

  #34269 commented on Jul 27, 2021 • 1 new comment

• Problem in adjust gamma function (tf.image.adjust_gamma)

  #50800 commented on Jul 27, 2021 • 1 new comment

• GazeML upgrade problem TF 1.4.0 to TF 2.5 (tf.contrib.layers.batch_norm)

  #50850 commented on Jul 27, 2021 • 1 new comment

• Address missing TensorFlow operations to TFLite:

  #21526 commented on Jul 27, 2021 • 1 new comment

• tensorflow.python.data.ops.dataset_ops.BatchDataset to HDF5

  #50840 commented on Jul 27, 2021 • 1 new comment

• tflite_runtime got installed successfully, however getting error while importing interpreter.

  #44944 commented on Jul 27, 2021 • 1 new comment

• Using the function signatures loaded from a SavedModel requires the original trackable object be kept in scope.

  #46708 commented on Jul 27, 2021 • 1 new comment

• Why is it so slow?

  #50718 commented on Jul 27, 2021 • 1 new comment

• cropping layer additional error message

  #50612 commented on Jul 28, 2021 • 1 new comment

• File system scheme 's3' not implemented

  #40302 commented on Jul 28, 2021 • 1 new comment

• copies data from gpu to cpu

  #50869 commented on Jul 28, 2021 • 1 new comment

• Pybind11 exception with tensorflow 2.2 env in python cpp communication

  #42405 commented on Jul 28, 2021 • 1 new comment

• Resource exhausted: MemoryError: Unable to allocate

  #45238 commented on Jul 28, 2021 • 1 new comment

• Move EthosU custom op out of AllOpsResolver

  #44628 commented on Jul 28, 2021 • 1 new comment

• Confine Tensorflow v2.3.2 C API to use only one thread.

  #47870 commented on Jul 28, 2021 • 1 new comment

• TFLite: Did not get operators, tensors, or buffers in subgraph 0. Error running on Android

  #45313 commented on Jul 28, 2021 • 1 new comment

• Unable to restore a layer of class TextVectorization - Text Classification

  #45231 commented on Jul 28, 2021 • 1 new comment

• Convert Tensorflow model to TFLite model - problem with boxes after quantization

  #44610 commented on Jul 28, 2021 • 1 new comment

• Same code - difference result from PC and jetson nano

  #45483 commented on Jul 28, 2021 • 1 new comment

• Linking issue: undefined reference to `tflite::micro::GetTensorShape SparkfunEdge

  #45927 commented on Jul 28, 2021 • 1 new comment

• TFlite interpreter stuck after call to invoke()

  #46107 commented on Jul 28, 2021 • 1 new comment

• Which version of CUDA and cuDNN should i use with tensorflow2.5.0 ?

  #50819 commented on Jul 28, 2021 • 1 new comment

• [TFLite] Fully Connected layers are not outputting accurate values with INT16 inputs of rank > 2

  #48959 commented on Jul 28, 2021 • 1 new comment

• device_lib.list_local_devices() InvalidArgumentError: Invalid device ordinal value (1). Valid range is [0, 0].

  #32793 commented on Jul 28, 2021 • 1 new comment

• tf.function runs correctly in eager mode, but breaks in graph mode

  #50832 commented on Jul 28, 2021 • 1 new comment

• Raspberry Pi 4 camera feed doesn't appear

  #45470 commented on Jul 28, 2021 • 1 new comment

• Keras model compiled with custom loss raises "Cannot convert a symbolic Keras input/output to a numpy array" error.

  #50678 commented on Jul 28, 2021 • 1 new comment

• Saved Keras model cannot be loaded using hub.KerasLayer API

  #50433 commented on Jul 28, 2021 • 1 new comment

• Prediction depends on Batch size

  #50860 commented on Jul 28, 2021 • 1 new comment

• Quantized Version of tf-lite model returning "ERROR: Didn't find op for builtin opcode 'CONV_2D' version '5'" when measuring performance

  #50652 commented on Jul 28, 2021 • 1 new comment

• tf.nn.local_response_normalization outputs NaN

  #46996 commented on Jul 28, 2021 • 1 new comment

• when i use tf2.4, tf.image_resize takes more time than tf1.2

  #50779 commented on Jul 29, 2021 • 1 new comment

• TF 2.5 | preprocessing.CategoryCrossing behavior different than feature_column.crossed_column

  #50821 commented on Jul 29, 2021 • 1 new comment

• TF -> TFLite conversion error: could not rewrite use of immutable bound input

  #50885 commented on Jul 29, 2021 • 1 new comment

• AttributeError: '_UserObject' object has no attribute 'add_slot'

  #50659 commented on Jul 29, 2021 • 1 new comment

• CMAKE ERROR for Android_arm64 of TFLite on MacOS.

  #50577 commented on Jul 29, 2021 • 1 new comment

• tf.cast(48584629,tf.float32) 精度丢失

  #50888 commented on Jul 29, 2021 • 1 new comment

• Hard to use a lstm tflite model.

  #46983 commented on Jul 29, 2021 • 1 new comment

• How the quantization on BERT

  #46565 commented on Jul 29, 2021 • 1 new comment

• GPU-delegate null EGLDisplay: lack of Wayland support?

  #46163 commented on Jul 29, 2021 • 1 new comment

• tf.experimental.numpy fails with plt.hist

  #46274 commented on Jul 29, 2021 • 1 new comment

• How to use the TF_VARIANT with C APIs?

  #50607 commented on Jul 29, 2021 • 1 new comment

• Core dumped when invoking TFLite model converted using latest nightly TFLite converter (2.4.0dev2020929)

  #43661 commented on Jul 29, 2021 • 1 new comment

• Unclear how to adapt multiple Keras preprocessing layers

  #50820 commented on Jul 29, 2021 • 1 new comment

• resnet50 inference result is wrong with tensorflow+oneDNN

  #50685 commented on Jul 29, 2021 • 1 new comment

• tf.keras.models.save_model not saving the probabilistic_model

  #50774 commented on Jul 29, 2021 • 1 new comment

• Quantization Aware Training with tf.GradientTape gives Error in TensorFlow2.x

  #41320 commented on Jul 29, 2021 • 1 new comment

• tf.keras cannot weight classes when using multiple outputs

  #41448 commented on Jul 29, 2021 • 1 new comment

• ESP32 doenst run correct anymore - Person Detection with Esp32 Camera

  #40956 commented on Jul 29, 2021 • 1 new comment

• The word vector obtained by the word2vec tutorial is very bad

  #50645 commented on Jul 29, 2021 • 1 new comment

• ResourceVariable GC bug

  #50796 commented on Jul 29, 2021 • 1 new comment

• Adding a feature of "Hard" attention could be very useful

  #50474 commented on Jul 29, 2021 • 1 new comment

• tf.math.sign() gives inconsistent output for the same input values for small complex numbers

  #50883 commented on Jul 30, 2021 • 1 new comment

• tflite runtime error with depthwise conv2D

  #26806 commented on Jul 30, 2021 • 1 new comment

• Problems when compiling with Clang

  #35939 commented on Jul 30, 2021 • 1 new comment

• [RNN] LSTM and Bidir layers can't be converted in a TFLite model

  #40510 commented on Jul 30, 2021 • 1 new comment

• Error Converting ssd_mobilenet_v2_oid_v4_2018_12_12 To Tflite

  #41117 commented on Jul 30, 2021 • 1 new comment

• Converting RESIZE_NEAREST_NEIGHBOR

  #41085 commented on Jul 30, 2021 • 1 new comment

• Multi-GPU issue: second A40 GPU returns zeros

  #50616 commented on Jul 30, 2021 • 1 new comment

• OP_REQUIRES failed at constant_op.cc

  #44428 commented on Jul 30, 2021 • 1 new comment

• WARNING:tensorflow:AutoGraph could not transform <function Model.make_predict_function.<locals>.predict_function at 0x000001FE83356310> and will run it as-is.

  #50894 commented on Jul 30, 2021 • 1 new comment

• Error when running inference on CPU using C++ bindings: Not found: Container localhost does not exist

  #41226 commented on Jul 31, 2021 • 1 new comment

• ParseTensor (tf.io.parse_tensor) is not vectorized - Vectorizing via tf.vectorized_map uses while_loop

  #43706 commented on Jul 31, 2021 • 1 new comment

• TFLite Conversion Error: Element Shape Required to be 1D

  #50942 commented on Jul 31, 2021 • 1 new comment

• Import error when multiple versions of tensorflow are in python path

  #50789 commented on Jul 31, 2021 • 1 new comment

• error: aliases are not supported on darwin In MacOS Big Sur

  #48130 commented on Jul 31, 2021 • 1 new comment

• tensorflow does not recognize gpu

  #50842 commented on Jul 31, 2021 • 1 new comment

• High memory consumption with model.fit in TF 2.x

  #40942 commented on Aug 1, 2021 • 1 new comment

• AttributeError: type object 'IteratorBase' has no attribute 'from_structure'

  #50871 commented on Aug 1, 2021 • 1 new comment

• TPU InternalError with TF 2.1.0 on Google Colab (Assigned device does not have registered OpKernel support for _Arg node iteratorgetnext_iterator)

  #35785 commented on Aug 1, 2021 • 1 new comment

• Error when running gmake for Tensorflow Lite Micro

  #35759 commented on Aug 1, 2021 • 1 new comment

• Unable to build hello_world for sparkfun_edge

  #35592 commented on Aug 1, 2021 • 1 new comment

• Tensorflow2.0 distributed training gives error :- A non-DistributedValues value 8 cannot be reduced with the given reduce op ReduceOp.SUM.

  #31852 commented on Aug 1, 2021 • 1 new comment

• 3D CNN with tf.nn.conv3d_transpose on CPU: 50-100 GB RAM free but segfault + "terminate called after throwing an instance of 'std::bad_alloc' what(): std::bad_alloc"

  #28313 commented on Aug 1, 2021 • 1 new comment

• No improvement in performance of deeplabv3_257_mv_gpu.tflite on TFLite for GPU

  #27613 commented on Aug 1, 2021 • 1 new comment

• Can LSTM be fully quantized for inference?

  #25563 commented on Aug 1, 2021 • 1 new comment

• Is tensorflow Java interface support Mac M1 slices ? Problematic frame: C [libtensorflow_framework.2.dylib+0x14c15] tensorflow::monitoring::MetricDef

  #48036 commented on Aug 1, 2021 • 1 new comment

• STM32F7 Hello World example fails for arm_cmplx_mag_squared_q10p6.c

  #35889 commented on Aug 1, 2021 • 1 new comment

• [TF2] TRT Engine Ops are not garbage collected, resulting in incorrect reuse

  #36675 commented on Aug 1, 2021 • 1 new comment

• MultiWorkerMirroredStrategy failed to run after a few batches occasionally

  #36953 commented on Aug 1, 2021 • 1 new comment

• RNN unrolled, wrong quantization graph

  #37023 commented on Aug 1, 2021 • 1 new comment

• Concat op not quantized

  #37544 commented on Aug 1, 2021 • 1 new comment

• grpc+verbs fail with tensorflow/core/common_runtime/process_state.cc:128] Check failed: 0 == cpu_allocators_.size() (0 vs. 1)

  #37622 commented on Aug 1, 2021 • 1 new comment

• Clarity issues in the mixed-precision execution guide

  #50870 commented on Aug 1, 2021 • 1 new comment

• Tensorflow could not build with `config=asan` flag

  #50892 commented on Aug 2, 2021 • 1 new comment

• preprocessor definition clash with glog

  #7480 commented on Aug 2, 2021 • 1 new comment

• TRT：converted tensorRT model pb too large

  #47547 commented on Aug 2, 2021 • 1 new comment

• Added ResNext models

  #37146 commented on Jul 30, 2021 • 1 new comment

• complex support for tf.math.argmin and tf.math.argmax

  #41441 commented on Jul 30, 2021 • 1 new comment

• support SparseToDenseOp on GPU

  #44702 commented on Jul 27, 2021 • 1 new comment

• [INTEL MKL]Register Eigen bf16 kernels with NoOps to make bf16 kernels(which In clear and infer list) can go through auto-mixed-precision pass.

  #44927 commented on Jul 28, 2021 • 1 new comment

• [Intel MKL] Fuse group of primitive ops for batch normalization to Fu…

  #44950 commented on Jul 28, 2021 • 1 new comment

• [INTEL MKL] Convert necessary PlaceholderWithDefault or Placeholder to Constant f…

  #45822 commented on Jul 28, 2021 • 1 new comment

• Lower ResizeNearestNeighbor when half_pixel_centers=true

  #46836 commented on Aug 1, 2021 • 1 new comment

• TFLu: Add UInt8 support for logistic operator

  #46873 commented on Jul 29, 2021 • 1 new comment

• Modernize TF_CPP_VMODULE handling using absl

  #47169 commented on Aug 1, 2021 • 1 new comment

• Addition of TanhExp activation function - issue #45929

  #47566 commented on Jul 29, 2021 • 1 new comment

• Add pre-load CPU feature guard check

  #48024 commented on Jul 29, 2021 • 1 new comment

• Map XLA opaque type to a scalar TensorShape in XLAShapeToTensorShape.

  #48233 commented on Jul 29, 2021 • 1 new comment

• Issue 47269: Disallow inline optimizer to optimize away function when it has been decorated with 'experimental_implements'

  #48856 commented on Jul 29, 2021 • 1 new comment

• [tf.data] graduate rejection_resample API from experimental to tf.data.Dataset

  #48894 commented on Jul 29, 2021 • 1 new comment

• Compatibility with NumPy 1.20

  #48935 commented on Jul 29, 2021 • 1 new comment

• Half pixel centers

  #49042 commented on Jul 29, 2021 • 1 new comment

• Made rounding in convert_image_dtype for numbers close to zero

  #49868 commented on Jul 29, 2021 • 1 new comment

• Allow `keras.utils.Sequence` sub-classes to use sparse/ragged tensors

  #49872 commented on Jul 29, 2021 • 1 new comment

• Fix swapping of tensor_content when loading a SavedModel on s390x

  #50152 commented on Jul 29, 2021 • 1 new comment

• Make DataFormatVecPermute's validation and implementation consistent with each other

  #50263 commented on Jul 29, 2021 • 1 new comment

• CombinedNMS plugin deactivated with TRT < 8. Only FP32 allowed with TRT >= 8

  #50450 commented on Jul 29, 2021 • 1 new comment

• [PluggableDevice] register default device for wrapdataset ops

  #50580 commented on Jul 29, 2021 • 1 new comment

• [PluggableDevice] add DEVICE_DEFAULT for python ops

  #50602 commented on Jul 29, 2021 • 1 new comment

• [PluggableDevice] add DEVICE_DEFAULT for batch_kernel/concat/fact/partioned_function ops

  #50603 commented on Jul 29, 2021 • 1 new comment

• [PluggableDevice] add DEVICE_DEFAULT for map_stage ops

  #50604 commented on Jul 29, 2021 • 1 new comment

• [PluggableDevice] add DEVICE_DEFAULT for iterator/optional ops

  #50605 commented on Jul 29, 2021 • 1 new comment

• [PluggableDevice] add DEVICE_DEFAULT for scopedallocator/shape ops

  #50608 commented on Jul 29, 2021 • 1 new comment

• [PluggableDevice] add DEVICE_DEFAULT for collective/bcast ops

  #50609 commented on Jul 29, 2021 • 1 new comment

• [PluggableDevice] add DEVICE_DEFAULT for debug/stage ops

  #50610 commented on Jul 29, 2021 • 1 new comment

• Refactor SparseConcat CPU implem and add two tests

  #50732 commented on Jul 27, 2021 • 1 new comment

• Support per-channel quantized INT8 weights unpacking in XNNPACK delegate

  #50875 commented on Aug 2, 2021 • 1 new comment

• Fix unchecked int32 indexing in some GPU kernels

  #50917 commented on Jul 30, 2021 • 1 new comment

• tf.profiler.experimental.Trace only works with name is "TraceContext"

  #50440 commented on Jul 26, 2021 • 0 new comments

• Grappler error when Softmax input has a variable dimension

  #50575 commented on Jul 26, 2021 • 0 new comments

• Feature Request: GPUOptions for Go binding

  #22926 commented on Jul 26, 2021 • 0 new comments

• Segmentation fault when using NCCL in MultiWorkerMirroredStrategy

  #50926 commented on Jul 27, 2021 • 0 new comments

• How to build tflite C++ library for ios from tf 2.3 onwards?

  #43955 commented on Jul 27, 2021 • 0 new comments

• Reappeared bug: TypeError: Parameter to MergeFrom() must be instance of same class: expected tensorflow.TensorShapeProto got tensorflow.TensorShapeProto.

  #50545 commented on Jul 27, 2021 • 0 new comments

• Did not get operators or tensors in subgraph 1. when using tf.lite.Interpreter

  #45444 commented on Jul 28, 2021 • 0 new comments

• Cannot convert model: Exception has occurred: ValueError

  #45460 commented on Jul 28, 2021 • 0 new comments

• implement reduce for the bitwise operations.

  #50939 commented on Jul 28, 2021 • 0 new comments

• Export Graph gradient API on Windows

  #50927 commented on Jul 28, 2021 • 0 new comments

• Error converting mask_rcnn/inception_resnet_v2_1024x1024 Tensorflow model from TFHub to Tflite

  #50750 commented on Jul 28, 2021 • 0 new comments

• tf.repeat memory consumption and performance

  #50712 commented on Jul 28, 2021 • 0 new comments

• Duplicate classes (com.google.flatbuffers) when flatbuffers is already in project

  #46356 commented on Jul 28, 2021 • 0 new comments

• MLIR converter quantizes conv2d before bias when dilaition rate > 1

  #50747 commented on Jul 28, 2021 • 0 new comments

• 48584629.0这个浮点数数据类型转换精度丢失

  #50889 commented on Jul 29, 2021 • 0 new comments

• There should be No expansion ReLU on MobileNet v2 bottleneck

  #50835 commented on Jul 29, 2021 • 0 new comments

• Trying to install tensorflow==2.0.3 with bazel on google app engine wont work

  #50557 commented on Jul 29, 2021 • 0 new comments

• TFLite conversion of LSTM model does not work with multiple batch size

  #50598 commented on Jul 29, 2021 • 0 new comments

• Error while building tensorflow from source.

  #50829 commented on Jul 29, 2021 • 0 new comments

• tf.compat.v1.wrap_function throws error when creating variable in tf.distribute scope

  #33585 commented on Jul 30, 2021 • 0 new comments

• Dataset.from_generator() returns a Dataset which freezes execution upon iteration

  #43768 commented on Jul 30, 2021 • 0 new comments

• Allow growth seems to take more memory than needed

  #29635 commented on Jul 30, 2021 • 0 new comments

• Tensorflow lite NnApiDelegate crashes on Pixel 3a XL

  #36645 commented on Aug 1, 2021 • 0 new comments

• Generalising class weights for all label ranks

  #46189 commented on Jul 29, 2021 • 0 new comments

• [INTEL MKL] Refactor mkl_relu_ops files, and add accuracy and performace UT.

  #46457 commented on Jul 29, 2021 • 0 new comments

• Get path from environment CUDA_PATH

  #46802 commented on Jul 29, 2021 • 0 new comments

• Custom os.path.join that is aware of TF filesystems

  #48125 commented on Jul 29, 2021 • 0 new comments

• [TFTRT] ConvertSplit for dynamic shapes.

  #48246 commented on Jul 29, 2021 • 0 new comments

• Add elastic training for workers in PSv2

  #48601 commented on Aug 1, 2021 • 0 new comments

• When events are specified as root events by user, they should be treated as root events in trace view.

  #48922 commented on Jul 29, 2021 • 0 new comments

• Update ICU to 69.1

  #49026 commented on Jul 27, 2021 • 0 new comments

• Remove duplicated Env::Default to fix gcs issue on Windows

  #49520 commented on Jul 30, 2021 • 0 new comments

• XLA persistent compilation cache

  #50071 commented on Jul 29, 2021 • 0 new comments

• [INTEL MKL] Comparison and cast op fusion

  #50113 commented on Aug 1, 2021 • 0 new comments

• [tf.data] graduate OptimizationOptions from experimental to tf.data

  #50279 commented on Jul 27, 2021 • 0 new comments

• Restore original framebuffer in ForceSyncTurning

  #50328 commented on Jul 29, 2021 • 0 new comments

• Extend the algorithm selector to consider a list of allowed and disallowed tactics

  #50408 commented on Jul 29, 2021 • 0 new comments

• [TFLite] Quantized int8/16 support for TILE operator

  #50424 commented on Jul 30, 2021 • 0 new comments

• Fix TransferLiteralFromDevice bug which appears in Tensorflow Master

  #50485 commented on Jul 27, 2021 • 0 new comments

• [MLIR][DISC] convert memref::alloc and memref::dealloc to ral.dispatch according to address space

  #50537 commented on Jul 26, 2021 • 0 new comments

• [PluggableDevice]add default device for function ops

  #50578 commented on Jul 29, 2021 • 0 new comments

• [PluggableDevice] register default devices for functional ops

  #50581 commented on Jul 29, 2021 • 0 new comments

• [PluggableDevice] add DEVICE_DEFAULT for datasets ops

  #50600 commented on Jul 29, 2021 • 0 new comments

• [PluggableDevice] add DEVICE_DEFAULT for Mutex/DestroyResource/Sequence ops

  #50601 commented on Jul 29, 2021 • 0 new comments

• Update errors_impl.py

  #50672 commented on Jul 29, 2021 • 0 new comments

• Changes to add FusedMatMul + Sigmoid fusion

  #50682 commented on Aug 1, 2021 • 0 new comments

• Fix the shape inference function for FusedBatchNormGradEx

  #50707 commented on Jul 30, 2021 • 0 new comments

• Documentation issue read me.

  #50729 commented on Jul 29, 2021 • 0 new comments

• [TFLite] Add int8 and int16x8 support for the ROUND operator

  #50786 commented on Jul 29, 2021 • 0 new comments

• FIx random gamma and space to depth tests.

  #50813 commented on Jul 30, 2021 • 0 new comments

• Update .bazelrc

  #50815 commented on Aug 1, 2021 • 0 new comments

• WIP: Implementation of a Buffer Packing Pass.

  #50834 commented on Jul 30, 2021 • 0 new comments

• Backport 'Make element_shape_except_first_dim as a local variable instead of in…' to r2.4

  #50877 commented on Jul 27, 2021 • 0 new comments

• [TF-TRT] Refactoring Top Level __init__ file for TF-TRT

  #50884 commented on Jul 29, 2021 • 0 new comments

• Mention intro-to-graphs in tf.function docstring.

  #50913 commented on Jul 28, 2021 • 0 new comments

• WIP: [XLA:GPU] BatchNorm lowering overhaul for cudnn fast fp16 kernels and act+side input fusion

  #50914 commented on Jul 29, 2021 • 0 new comments

• [PluggableDevice] fix While segment fault on PluggableDevice

  #50916 commented on Jul 29, 2021 • 0 new comments

• [Go] add bazel rules

  #50934 commented on Jul 29, 2021 • 0 new comments