tensorflow / tensorflow

30 Pull requests merged by 22 people

• [ROCm] Fix for ROCM CSB breakage - 201007

  #43852 merged Oct 14, 2020

• Add Neg gradient

  #43935 merged Oct 14, 2020

• Official libtensorflow links

  #43980 merged Oct 14, 2020

• [Intel MKL] Fixing mkl_eager_op_rewrite_test error

  #43981 merged Oct 14, 2020

• Enforce that the target makefile must be called TARGET_makefile.

  #43896 merged Oct 13, 2020

• Bump junit from 4.11 to 4.13.1 in /tensorflow/java/maven/tensorflow-hadoop

  #43971 merged Oct 13, 2020

• Bump junit from 4.11 to 4.13.1 in /tensorflow/java/maven/spark-tensorflow-connector

  #43970 merged Oct 13, 2020

• Cleanup Selectv2 broadcasting

  #43907 merged Oct 13, 2020

• [TF:MLIR] Add TF to TF lowering patterns to legalize_tf pass for TFLite

  #42525 merged Oct 13, 2020

• [INTEL MKL] Enable DNNL BatchMatMul support with broadcast and update oneDNN to v1.6.4.

  #43741 merged Oct 12, 2020

• [Intel MKL] Adding support for MaxPool with native format

  #43506 merged Oct 12, 2020

• Fix for concat v2 unit test failure in [INTEL MKL] enabled Tensorflow

  #43573 merged Oct 9, 2020

• [INTEL MKL] Bug-fix to in-place computation with tensor forwarding.

  #43577 merged Oct 9, 2020

• [Intel MKL] Enabling FusedMatMul op with native format

  #43658 merged Oct 9, 2020

• Split env library so that there is only one instance of `Env::Default()`

  #43594 merged Oct 9, 2020

• Patch kissfft to remove sys/types.h include from kiss_fft.h.

  #43867 merged Oct 9, 2020

• Add oneDNN-aarch64 reference build option

  #43889 merged Oct 9, 2020

• _compute_causal_padding() missing 1 required positional argument: 'inputs'

  #42466 merged Oct 9, 2020

• fix keras progbar output collision (2)

  #43873 merged Oct 9, 2020

• [Intel MKL] Adding support for FusedBatchNorm op with native format

  #43504 merged Oct 8, 2020

• Add GPU kernel for ApplyProximalAdagrad

  #43814 merged Oct 8, 2020

• Alternate approach (to PR #43240) to fixing ethos-u build errors.

  #43858 merged Oct 8, 2020

• Fix TFLite quantization example

  #43754 merged Oct 8, 2020

• Update flatbuffers dependency

  #43856 merged Oct 8, 2020

• Autotune Dataset.map when transforming builtin map with autograph

  #43859 merged Oct 8, 2020

• Update nn_ops.py

  #42324 merged Oct 8, 2020

• Update tensorflow.keras.metrics.pbtxt

  #42442 merged Oct 8, 2020

• [Intel MKL] Adding support for AvgPool with native format

  #43505 merged Oct 8, 2020

• [XLA/GPU] clear operands for removed HLOs.

  #43746 merged Oct 8, 2020

• Set the "nightly" branch to master HEAD every day

  #43803 merged Oct 8, 2020

31 Pull requests opened by 25 people

• Use sysctl for maximum CPU frequency on MacOS

  #43855 opened Oct 7, 2020

• Fix tf.function speacial casing for forwardporp (placeholder issue)

  #43864 opened Oct 8, 2020

• Refactor SparseApplyFtrl CPU kernel into class

  #43868 opened Oct 8, 2020

• remove error msg if >1 adapter can handle data (2)

  #43874 opened Oct 8, 2020

• Fix channel first support for effieicentnet

  #43880 opened Oct 8, 2020

• Update Keras 2D convolution docs

  #43884 opened Oct 8, 2020

• [XLA] When ptxas do not know about an SM, fallback to the driver.

  #43888 opened Oct 8, 2020

• [INTEL MKL] MKL DNN 0.x code cleanup - Quantized Matmul op

  #43895 opened Oct 8, 2020

• [TF-TRT] Cast now supports more conversions if TRT >= 7.0.0.11

  #43900 opened Oct 8, 2020

• Allow custom image size for NASNetMobile

  #43918 opened Oct 9, 2020

• [INTEL MKL] Added bf16 support for Sparse_Xent_op.

  #43920 opened Oct 9, 2020

• Compare Update Pull

  #43921 opened Oct 10, 2020

• Add additional example to timeseries_dataset_from_array

  #43930 opened Oct 10, 2020

• Use TensorShape.assert_is_compatible_with instead of merge_with

  #43932 opened Oct 10, 2020

• VLOG(1) an error if cuDNN sub-lib pre-initialization fails

  #43949 opened Oct 12, 2020

• Print the version of the cuDNN library found at runtime

  #43950 opened Oct 12, 2020

• Avoid empty linker_bin_path breaking the build

  #43951 opened Oct 12, 2020

• [TFLite] Add int16x8 support for RESIZE_BILINEAR operator

  #43957 opened Oct 12, 2020

• Horizontal input fusion again

  #43964 opened Oct 13, 2020

• DOC Add None as possible in input_shape

  #43975 opened Oct 13, 2020

• Bump junit from 4.11 to 4.13.1.

  #43987 opened Oct 13, 2020

• Bump junit from 4.11 to 4.13.1.

  #43988 opened Oct 13, 2020

• Bump junit from 4.11 to 4.13.1.

  #43989 opened Oct 13, 2020

• Bump junit from 4.11 to 4.13.1.

  #43990 opened Oct 13, 2020

• Bump junit from 4.11 to 4.13.1.

  #43991 opened Oct 13, 2020

• fixing problem with patching CMSIS lib when installing under msys64.

  #44001 opened Oct 14, 2020

• Update download_and_extract script for faster CMSIS patching

  #44014 opened Oct 14, 2020

• Add correct scatter file values for STM32F4 target

  #44015 opened Oct 14, 2020

• [PluggableDevice] Kernel C API enhancement for retrieving attributes

  #44017 opened Oct 14, 2020

• Update README.md

  #44019 opened Oct 14, 2020

• [MLIR] Fix verifier for TF::Conv2DOp and TF::Conv3DOp.

  #44022 opened Oct 14, 2020

103 Issues closed by 59 people

• Exploding loss when converting tf1 to tf2

  #43809 closed Oct 14, 2020

• i can't find tutorials. Examples.

  #43602 closed Oct 14, 2020

• Connecting to invalid output 1 of source node PartitionedCall_7 which has 1 outputs

  #43942 closed Oct 14, 2020

• which cuda version is required for the pre-built c library

  #43908 closed Oct 14, 2020

• To compile the code need portpicker. What for? is it safe?

  #43926 closed Oct 14, 2020

• Could not find any cudnn.h matching version '8.0.4' in any subdirectory

  #43956 closed Oct 14, 2020

• How to print the tensor values in tensorflow 2.x version?

  #43993 closed Oct 14, 2020

• TF to TF Lite Int 8 resulting in error:

  #43984 closed Oct 14, 2020

• Why TF loss function gives different results than Pytorch?

  #43997 closed Oct 14, 2020

• invalid syntax error while importing tensorflow and keras

  #43419 closed Oct 14, 2020

• Error "Cannot iterate over a tensor with unknown first dimension" when using TF Keras Attention Layer.

  #43651 closed Oct 14, 2020

• android use XNNPACK happen error

  #42056 closed Oct 14, 2020

• Error in tflite model deployment on android：Cannot copy to a TensorFlowLite tensor with 270000 bytes from a Java Buffer with 1228800 bytes.

  #43965 closed Oct 14, 2020

• How to use custom tensorflow op for CUDA?

  #43994 closed Oct 14, 2020

• RuntimeError: Attempting to use a delegate that only supports static-sized tensors with a graph that has dynamic-sized tensors.Failed to apply the default TensorFlow Lite delegate indexed at 0.

  #43569 closed Oct 14, 2020

• TRTEngineOp is not used

  #43644 closed Oct 14, 2020

• Add Post-training integer quantization converter for depth_to_space

  #38626 closed Oct 13, 2020

• tf.io.gfile.listdir does not with with S3 buckets

  #36893 closed Oct 13, 2020

• Tensorflow session.run command 20% slower in tf2.3 compared to tf1.15

  #43620 closed Oct 13, 2020

• Tensorflow 2.3.0 stopped detecting GPU suddenly while it used to detect in past

  #43785 closed Oct 13, 2020

• Problem loading model trained with EfficientNetB0

  #43576 closed Oct 13, 2020

• TPU V3-1024 training crashes randomly

  #43707 closed Oct 13, 2020

• Memory blowup converting from TFLite to C

  #43749 closed Oct 13, 2020

• Segmentation fault: 11

  #41838 closed Oct 13, 2020

• Inconsistent keras.layers.normalization version

  #43871 closed Oct 13, 2020

• bug in Using the SavedModel format Guide

  #41679 closed Oct 13, 2020

• BatchNorm generates NaN moving_variance on GPU with fused set to True for some inputs

  #34062 closed Oct 13, 2020

• all_ops_resolver.h:4:10: fatal error: tensorflow/lite/micro/compatibility.h: No such file or directory

  #43881 closed Oct 13, 2020

• tf.contrib.lookup.string_to_index_table_from_file optimization

  #43351 closed Oct 13, 2020

• Shape information is lost with DepthwiseConv2D

  #37039 closed Oct 13, 2020

• 'Operation' object has no attribute '_unconditional_update' when using pre-trained model that used batchnorm

  #43552 closed Oct 13, 2020

• GPU support Software requirements links to incompatible CUDA Toolkit, cuDNN version

  #43933 closed Oct 13, 2020

• Error in loading tflite model at runtime from native C++ library.

  #41619 closed Oct 12, 2020

• Fail to load model from ModelCheckpoint in java.

  #42060 closed Oct 12, 2020

• Import issue with tensor flow Windows 10 python 3.7

  #43750 closed Oct 12, 2020

• The parameters for BatchNormalization are not updated

  #40249 closed Oct 12, 2020

• Ensemble of custom sequential models

  #38224 closed Oct 12, 2020

• Custom Op Output Shape is Unknown in TF 1.15

  #43897 closed Oct 12, 2020

• data_element may be type tuple, instead of tf.Tensor

  #39377 closed Oct 12, 2020

• nccl_ops.all_sum does not correctly reduce gradients

  #41539 closed Oct 12, 2020

• Model trains on GPU but not on TPU

  #43276 closed Oct 12, 2020

• from tensorflow.contrib.seq2seq import Helper

  #43612 closed Oct 12, 2020

• Bad Loss History, Callback vs. Progressbar Loss Mismatch

  #43470 closed Oct 12, 2020

• Error. Converter does not support Quantization NN with 'tanh' activation

  #42390 closed Oct 12, 2020

• tf.keras custom layer does not use "compute_output_shape"

  #38296 closed Oct 12, 2020

• Unable to use FeatureColumn with Keras Functional API

  #27416 closed Oct 12, 2020

• Compiling client code with -mavx2 leads to Eigen errors

  #43666 closed Oct 12, 2020

• tf-nightly fails to initialize TPUs

  #43581 closed Oct 12, 2020

• [TF 2.2 API docs] tf.keras.applications.resnet_v2.preprocess_input docs Returns paragraph is misleading

  #39677 closed Oct 12, 2020

• Lots of links are Broken.

  #37607 closed Oct 12, 2020

• Batch normalisation with shared vision model sometimes causes NaN.

  #43603 closed Oct 12, 2020

• Some times ,interpreter_->Invoke() doesn't return

  #43600 closed Oct 12, 2020

• ImportError: DLL load failed: A dynamic link library (DLL) initialization routine failed.

  #43579 closed Oct 12, 2020

• PyCharms: Unresolved reference 'config' for some of the TensorFlow imports. For example: from tensorflow.python.autograph.core import config from tensorflow.python.autograph.core import converter from tensorflow.python.autograph.impl import api

  #43952 closed Oct 12, 2020

• When compile keras model with multi losses, regularizers are not added properly

  #43953 closed Oct 12, 2020

• The format of the table is wrong.

  #43609 closed Oct 12, 2020

• Keras on_epoch_end() callback output happens BEFORE completion of epoch output with verbose=1

  #43184 closed Oct 12, 2020

• Unable to build tflite for aarch64

  #43811 closed Oct 12, 2020

• TFLITE not enable RUY AVX runtime path

  #43360 closed Oct 12, 2020

• RTX3080 install RuntimeError: CUDA runtime implicit initialization on GPU:0 failed. Status: device kernel image is invalid

  #43701 closed Oct 12, 2020

• GPU support does not install libcublas.so.10 on TF 2.3

  #43945 closed Oct 11, 2020

• Slow inference with libopencm3 when compared to Mbed

  #43931 closed Oct 11, 2020

• Kernel changes break xtensa_hifimini

  #43684 closed Oct 11, 2020

• Internal launch timed out and terminated

  #43914 closed Oct 10, 2020

• Tf 2.0 training performance issue

  #35542 closed Oct 10, 2020

• 2.3 Nightly build produces error when initializing system in TF documentation's Colab tutorial

  #40196 closed Oct 10, 2020

• RGB to YUV: Conflicting information between example and description

  #41496 closed Oct 10, 2020

• Docs: incomplete description of callbacks.ModelCheckpoint's save_best_only parameter

  #41420 closed Oct 10, 2020

• tf.where raises ValueError for RaggedTensor arguments

  #37351 closed Oct 9, 2020

• Using intermediate layer outputs in custom loss function causes CUDA_ERROR_OUT_OF_MEMORY

  #33837 closed Oct 9, 2020

• creating tf.data.Dataset object from generator gives error for incorrect file name

  #43757 closed Oct 9, 2020

• OperatorNotAllowedInGraphError: using a `tf.Tensor` as a Python `bool` is not allowed in Graph execution. Use Eager execution or decorate this function with @tf.function.

  #43876 closed Oct 9, 2020

• Keras way of auto-naming layers triggers a naming issues when loading an existing model, and slicing its output

  #43824 closed Oct 9, 2020

• OP_REQUIRES failed : Not found: No algorithm worked!

  #43223 closed Oct 9, 2020

• Converted model gives shape error during inference in C++ API.

  #43624 closed Oct 9, 2020

• Building micro_speech for Arm Mbed OS using ARM Compiler (armclang) fails

  #43848 closed Oct 9, 2020

• tf.numpy_function() does not support symbolic Tensors

  #43542 closed Oct 9, 2020

• tf.signal.stft has side effects

  #43909 closed Oct 9, 2020

• TF 2.1: ModelCheckpoint save best model fails

  #43540 closed Oct 9, 2020

• Warning when trying to Load CIFAR10 in Tensorflow 1.x

  #43429 closed Oct 9, 2020

• Training of a simple model with KerasLayer hangs on macOSX with tf 2.3.0

  #43060 closed Oct 9, 2020

• TFLite C++ API outputs the same results all the time

  #43788 closed Oct 9, 2020

• WARNING:tensorflow:A checkpoint was restored (e.g. tf.train.Checkpoint.restore or tf.keras.Model.load_weights) but not all checkpointed values were used. See above for specific issues.

  #43554 closed Oct 9, 2020

• when i am running the sample tutorial code of basic text classification i just cannot download a file named imdb_word_index.json

  #43245 closed Oct 9, 2020

• Confused which version TensorfFlow CPU supports

  #43229 closed Oct 9, 2020

• cuda and tensorflow

  #43352 closed Oct 9, 2020

• Installing Tensorflow

  #43479 closed Oct 9, 2020

• Issue on Load Trax

  #43086 closed Oct 9, 2020

• OSError: [WinError 126] The specified module could not be found even when following documentation

  #43776 closed Oct 9, 2020

• Error computing gradient of a vectorized scan loop

  #41789 closed Oct 8, 2020

• BinaryCrossentropy and binary_crossentropy in the. same `tf.keras.losses` module

  #43827 closed Oct 8, 2020

• Error downloading [https://mirror.bazel.build/github.com/unicode-org/icu/archive/release-62-1.tar.gz

  #43893 closed Oct 8, 2020

• Difference between keras and tf.keras

  #43520 closed Oct 8, 2020

• call back error

  #43501 closed Oct 8, 2020

• Tensorflow 2.3.0 can't detect CUDA on Python 3.7.9

  #43528 closed Oct 8, 2020

• imx8 tflite opencl

  #43063 closed Oct 8, 2020

• is it possible tensorflow-gpu on RK3399 mali t-860?

  #43807 closed Oct 8, 2020

• why does tensorflow import_pb_to_tensorboard.py tool require the pb file to be a fixed name "saved_model.pb"?

  #43825 closed Oct 8, 2020

• map_and_batch() is way faster than map().batch() ???

  #43489 closed Oct 8, 2020

• Issue when convert to tf lite

  #43854 closed Oct 8, 2020

• regression in layer names for tf-operation layers in current tf-nightly (2.4)

  #43840 closed Oct 7, 2020

• tf.feature_column.embedding_column: wrong dtype in lookup variable

  #43628 closed Oct 7, 2020

• tf.keras.backend.sqrt(tf.constant(-1.0)) is 0 which is misleading and tf.sqrt(tf.constant(-1.0)) is 'nan' which is the way it should be.

  #34165 closed Oct 7, 2020

96 Issues opened by 91 people

• Validation accuracy stuck at 0.6 in my CNN model

  #44024 opened Oct 14, 2020

• In TFnighty TFLite Interpreter missing setUseNNAPI "wrapper" method to call Interpreter.Options.setUseNNAPI

  #44023 opened Oct 14, 2020

• Different output for custom tf.keras.Model within tf.function

  #44021 opened Oct 14, 2020

• Problem to import keras from tensorflow 2.3.1

  #44020 opened Oct 14, 2020

• I've created a DCGAN model and when i run that in tensorflow ver 1.x and 2.x, it gives different result.What is the the reason for this?Thanks

  #44018 opened Oct 14, 2020

• TFLite not respecting ByteBuffer's limit

  #44016 opened Oct 14, 2020

• Patching CMSIS source file takes a lot of time

  #44013 opened Oct 14, 2020

• CUDA issues with 'tf-2.4.0-dev20201012'

  #44012 opened Oct 14, 2020

• tf.print doesn't work as expected across devices in for loop in graph mode

  #44011 opened Oct 14, 2020

• ImportError: cannot import name 'MomentumParameters'

  #44010 opened Oct 14, 2020

• ValueError: Unable to create group (Name already exists)

  #44009 opened Oct 14, 2020

• Using skip() of tf.data.Dataset is slow for big datasets

  #44008 opened Oct 14, 2020

• Output of Regression Quantized Aware Training Model is always restricted between 0 and 6

  #44007 opened Oct 14, 2020

• tf.keras.utils.Generator doesn't get called after first batch.

  #44006 opened Oct 14, 2020

• ValueError: The same saveable will be restored with two names: layer_with_weights-1/_table/.ATTRIBUTES/table

  #44005 opened Oct 14, 2020

• Cannot import Tensorflow due to DLL import error

  #44004 opened Oct 14, 2020

• segmentation fault using detection postprocess for ssd_mobilenet_v3

  #44003 opened Oct 14, 2020

• How to serve a model containing ```tensorflow.keras.layers.experimental.preprocessing``` layers ?

  #44002 opened Oct 14, 2020

• Comparison of gpu acceleration between gtx 1660ti(6g video memory) and quadro gv100(32g video memory)

  #44000 opened Oct 14, 2020

• Why is loading resnet50 keras model in TF 1.15 on TPU does not work?

  #43999 opened Oct 14, 2020

• Update BatchNormalization documentation

  #43998 opened Oct 14, 2020

• Build options

  #43996 opened Oct 14, 2020

• Regularization Loss is getting added twice if model is encapsulated in another model

  #43995 opened Oct 14, 2020

• Passing structs by value makes C API hard to wrap

  #43992 opened Oct 14, 2020

• installing tflite-model-maker encounters errors in regards to tf-nightly==2.4.0.dev20200810

  #43986 opened Oct 13, 2020

• TF to TF Lite Int 8 resulting in error: "Quantization not yet supported for op: %"

  #43985 opened Oct 13, 2020

• remote devices in cluster not visible in TF 2

  #43983 opened Oct 13, 2020

• AttributeError: module 'tensorflow._api.v2.sets' has no attribute 'set_intersection'

  #43982 opened Oct 13, 2020

• ERROR: An error occurred during the fetch of repository 'local_config_cc':

  #43979 opened Oct 13, 2020

• Cannot save mixed precision model when using activity regularizer

  #43978 opened Oct 13, 2020

• Method to get k smallest elements from a tensor

  #43977 opened Oct 13, 2020

• run-time error

  #43976 opened Oct 13, 2020

• about RuntimeError: Use `_distributed_apply()` instead of `apply_gradients()` in a cross-replica context.

  #43974 opened Oct 13, 2020

• Error when bazel-building tensorflow from source

  #43972 opened Oct 13, 2020

• Memory leak with tf.data.Dataset.map

  #43969 opened Oct 13, 2020

• Adam Optimizer not working in migration of custom model

  #43968 opened Oct 13, 2020

• could tf.feature_column.shared_embeddings support sequence_categorical_column_with_* in the future?

  #43967 opened Oct 13, 2020

• AssertionError: Could not compute output Tensor("ctc/Identity:0", shape=(None, 1), dtype=float32)

  #43966 opened Oct 13, 2020

• 'tf.Conv2D' op is neither a custom op nor a flex op error while converting ONNX model to TF Lite

  #43963 opened Oct 12, 2020

• Cannot use seq_len other than 128 with BertNLClassifier – MobileBERT Model Maker Text Classification example

  #43962 opened Oct 12, 2020

• Model Initialization take too much time

  #43961 opened Oct 12, 2020

• the number of parameters of a GRU layer is different on keras and tensorflow 2.X

  #43960 opened Oct 12, 2020

• After converting to TFLite the network lost structure and return zero

  #43959 opened Oct 12, 2020

• Enable profiling in release builds

  #43958 opened Oct 12, 2020

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

  #43955 opened Oct 12, 2020

• tf.keras.models.model_load outputs random evaluation values

  #43954 opened Oct 12, 2020

• tf.function and PerReplica tf.Variable does not work together.

  #43948 opened Oct 12, 2020

• CUDA 11.1 error on tf-nightly - libcusolver.so.10 not found

  #43947 opened Oct 12, 2020

• Tensorflow Import Error: DLL load failed....

  #43946 opened Oct 11, 2020

• Keras: removing layers by creating new model not working

  #43944 opened Oct 11, 2020

• Cannot load data from imdb_reviews datasets.

  #43943 opened Oct 11, 2020

• Masking scores before Softmax in Attention discussed here - https://www.tensorflow.org/tutorials/text/nmt_with_attention

  #43941 opened Oct 11, 2020

• support Quantization Aware Training for Conv2dTranspose->BathcNorm->Activation

  #43939 opened Oct 11, 2020

• [EfficientNetB7] validation not getting correct accuracy

  #43938 opened Oct 11, 2020

• please expose uniform_full_int on tf.random

  #43937 opened Oct 11, 2020

• In the custom model, keras.layers.Embedding Output Shape is multiple

  #43936 opened Oct 11, 2020

• unresolved custom op: FusedBatchNormV3.Node number 13 (FusedBatchNormV3)

  #43934 opened Oct 11, 2020

• Missing motivation for TF hacking in Pix2Pix Generative tutorial

  #43929 opened Oct 10, 2020

• Using Orthogonal initializer in dense layer with backend set to fp16 gets stuck

  #43928 opened Oct 10, 2020

• Docker Image : cuda:10.0-cudnn7-devel-ubuntu16.04 -Training performance low - GPU Util% keep changing - Process Id section empty

  #43927 opened Oct 10, 2020

• Avoiding data copy when using tf.py_func in Python

  #43925 opened Oct 10, 2020

• Build from source wheel can not be used in another cpu

  #43924 opened Oct 10, 2020

• Docker cuda:10.0-cudnn7-devel-ubuntu16.04 - 3 out of 4 GPUs not utilized

  #43919 opened Oct 9, 2020

• tensorflow/core/grappler/optimizers/meta_optimizer.cc:581

  #43917 opened Oct 9, 2020

• Keras Functional API does not support tf.custom_gradient() with tf.numpy_function()

  #43916 opened Oct 9, 2020

• ImportError: DLL load failed while importing _pywrap_tensorflow_internal: A dynamic link library (DLL) initialization routine failed.

  #43915 opened Oct 9, 2020

• keras.utils.Sequence instance can't be deleted from memory after training

  #43913 opened Oct 9, 2020

• Optimized cadence kernels need to use the new TfLiteEvalTensor API

  #43912 opened Oct 9, 2020

• "device kernel image is invalid" when using pre-built linux+gpu c library

  #43911 opened Oct 9, 2020

• Optimized arc kernels need to use the new TfLiteEvalTensor API

  #43910 opened Oct 9, 2020

• Prefetch to GPU: prefetch_to_device does not do anything (from_generator), tf.data.Dataset API unclear

  #43905 opened Oct 9, 2020

• Per Example Gradients fail with LSTM

  #43904 opened Oct 9, 2020

• Link error occurs due to MetaGraphDef

  #43903 opened Oct 9, 2020

• Supporting class_weight in train_step ?

  #43902 opened Oct 9, 2020

• TARGET command line option to the TFLM Makefile and the target-specific makefile have specific naming requirements

  #43898 opened Oct 8, 2020

• Error while building tensorflow from source, need it for C++ inference.

  #43894 opened Oct 8, 2020

• Model size reduction in set_weights()

  #43892 opened Oct 8, 2020

• No way to initialize Interpreter with Model reference in TensorFlowLiteSwift

  #43891 opened Oct 8, 2020

• pytorch sees my gpu, but tensorflow does not

  #43890 opened Oct 8, 2020

• Building on MacOS is broken for TFLite Micro

  #43887 opened Oct 8, 2020

• Update Keras Conv2D layer docs to specify padding value

  #43886 opened Oct 8, 2020

• Build Error with codes from TinyML Book

  #43885 opened Oct 8, 2020

• Embedding Projector - Unable to load different data set

  #43883 opened Oct 8, 2020

• folding batchnorm into conv in per-tensor weights quantization

  #43882 opened Oct 8, 2020

• Using SYSTEM cURL (build against OpenSSL) conflicts with BoringSSL

  #43878 opened Oct 8, 2020

• TFLite: Getting error "Expected bias tensor to be a vector" when trying to convert and quantize a model

  #43877 opened Oct 8, 2020

• Unable to load Cats vs Dogs dataset

  #43875 opened Oct 8, 2020

• [Encountered unresolved custom op: ResizeNearestNeighbor] when running tflite model

  #43872 opened Oct 8, 2020

• Bump docker nightly cuda version to 11.1

  #43870 opened Oct 8, 2020

• No zero values inside callback functions

  #43869 opened Oct 8, 2020

• got error while import keras libraries

  #43866 opened Oct 8, 2020

• Wrong linker flags in Makefile for ARM Compiler

  #43865 opened Oct 8, 2020

• Getting tensorflow.python.framework.errors_impl.NotFoundError

  #43863 opened Oct 8, 2020

• Reading cuda array in c++ as tensor and reading that tensor in python

  #43862 opened Oct 8, 2020

• Unable to Install TF in Conda

  #43861 opened Oct 8, 2020

• CUDA driver is not load in docker container

  #43860 opened Oct 8, 2020

248 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.

• Adding per-replica dataset distribution

  #38968 commented on Oct 14, 2020 • 48 new comments

• [XLA/GPU] Size-constrained buffer allocation.

  #43850 commented on Oct 14, 2020 • 39 new comments

• [INTEL MKL] Changes to remove MKL binary blob.

  #43586 commented on Oct 14, 2020 • 26 new comments

• [PluggableDevice] Add subdevice type support

  #43611 commented on Oct 14, 2020 • 25 new comments

• [PluggableDevice] Add load mechanism and streamexecutor C API extension

  #43610 commented on Oct 14, 2020 • 15 new comments

• TFlu: Add Cortex-M generic target makefile

  #43726 commented on Oct 14, 2020 • 14 new comments

• XLA Asynchronous compilation

  #43034 commented on Oct 14, 2020 • 10 new comments

• Use std::endl to force flush of output

  #43513 commented on Oct 14, 2020 • 10 new comments

• Large batch size with dense layers will fail all_reduce occasionally

  #41980 commented on Oct 12, 2020 • 9 new comments

• Support RESIZE_BILINEAR in TFLu

  #43426 commented on Oct 14, 2020 • 9 new comments

• relaxing numpy requirement

  #43679 commented on Oct 12, 2020 • 8 new comments

• [INTEL MKL] Fix control_flow_ops_test unit test failure

  #43798 commented on Oct 12, 2020 • 8 new comments

• java.lang.IllegalArgumentException: Internal error: Failed to apply delegate: Attempting to use a delegate that only supports static-sized tensors with a graph that has dynamic-sized tensors

  #38036 commented on Oct 8, 2020 • 7 new comments

• Tensorflow lite(ssd_mobilenet_v2_fpnlite_640*640) Android APP detect nothing

  #43583 commented on Oct 13, 2020 • 7 new comments

• Go installation error

  #43847 commented on Oct 14, 2020 • 7 new comments

• Converting a LSTM model from Pytorch to Tensorflow using ONNX

  #43808 commented on Oct 13, 2020 • 6 new comments

• Cannot confine TensorFlow C API to use not more than 1 threads in total

  #42510 commented on Oct 14, 2020 • 6 new comments

• Fix docs for model_from_config

  #40811 commented on Oct 14, 2020 • 6 new comments

• [INTEL MKL] MKL DNN 0.x code cleanup - FusedBatchNorm op

  #43548 commented on Oct 14, 2020 • 6 new comments

• fix TFOpLambda name for KerasOpDisparcher

  #43844 commented on Oct 13, 2020 • 6 new comments

• SavedModel does not work correctly with signatures that use structured inputs or outputs

  #40501 commented on Oct 9, 2020 • 5 new comments

• Interpreter takes too much time to return results while similar iOS app makes it 10 times faster

  #43787 commented on Oct 11, 2020 • 5 new comments

• Loss and metric calculated differently on validation

  #43763 commented on Oct 13, 2020 • 5 new comments

• Tensorflow 2.3 CUDNN - Detected cudnn out-of-bounds write in convolution buffer! + GPU Sync Failed

  #43623 commented on Oct 14, 2020 • 5 new comments

• UnknownError: Failed to get convolution algorithm.

  #43764 commented on Oct 14, 2020 • 5 new comments

• Add support lite model conversion with zero inputs

  #43558 commented on Oct 14, 2020 • 5 new comments

• [INTEL MKL] Adding CentOS 8 oneDNN partials and dockerfiles

  #43659 commented on Oct 14, 2020 • 5 new comments

• Expose TF_RegisterFilesystemPlugin C API

  #43680 commented on Oct 13, 2020 • 5 new comments

• OP_REQUIRES failed at sparse_tensor_dense_matmul_op.cc:146 : Invalid argument: k (784) from index[4352,1] out of bounds (>=784)

  #43842 commented on Oct 9, 2020 • 4 new comments

• NNAPI Delegate BUG

  #43541 commented on Oct 9, 2020 • 4 new comments

• Horovod fails to train Keras Preprocessing IntegerLookup Layer

  #43834 commented on Oct 10, 2020 • 4 new comments

• Tensorflow 2.2 and 2.3 not detecting GPU with CUDA 10.1

  #43236 commented on Oct 11, 2020 • 4 new comments

• TF 2.3.0 build failure when its conda package is built using Anaconda's toolchain

  #42313 commented on Oct 12, 2020 • 4 new comments

• AttributeError: 'Tensor' object has no attribute '_lazy_read' inside tf.while_loop containing tf.scatter_nd_update

  #43851 commented on Oct 13, 2020 • 4 new comments

• Debugger V2 not working. Invalid argument: DebugNumericSummaryV2Op requires tensor_id to be less than or equal to (2^53). Given tensor_id:26

  #43608 commented on Oct 13, 2020 • 4 new comments

• Enable multiple input pipelines with a single machine, MultiWorkerMirroredStrategy's experimental_distribute_datasets_from_function

  #43832 commented on Oct 14, 2020 • 4 new comments

• tf.keras.Sequential() fails

  #41855 commented on Oct 9, 2020 • 3 new comments

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

  #43768 commented on Oct 9, 2020 • 3 new comments

• how to save a model with weights and paramters

  #43643 commented on Oct 12, 2020 • 3 new comments

• Training AlexNet structure on entire ImageNet in Tensorflow 2.0, But why the loss and accuracy are all flat?

  #43503 commented on Oct 13, 2020 • 3 new comments

• Metal delegate Crash with C++ interface

  #43615 commented on Oct 14, 2020 • 3 new comments

• "IndexError: list index out of range" when load EfficientDet SavedModel with tf.keras

  #43527 commented on Oct 14, 2020 • 3 new comments

• Mutablehashtable lookup support full size dynamic default values.

  #43269 commented on Oct 14, 2020 • 3 new comments

• Fixes #43200 - TensorBoard histogram breaks on unsupported weight dtypes

  #43418 commented on Oct 13, 2020 • 3 new comments

• eigen download link points to google storage instead of bitbucket

  #43499 commented on Oct 14, 2020 • 3 new comments

• Updated documentation of TimeDistributed

  #43797 commented on Oct 14, 2020 • 3 new comments

• replace PFLR DeviceGetContext hardcode with Device::IsRemoteCallAllowed

  #43812 commented on Oct 14, 2020 • 3 new comments

• TF2.3 converter.convert ValueError: Input 0 of node StatefulPartitionedCall/functional_1/resnet50/conv1_bn/AssignNewValue was passed float from Func/StatefulPartitionedCall/input/_5:0 incompatible with expected resource.

  #43833 commented on Oct 8, 2020 • 2 new comments

• getting these errors in tensorflow.contrib, TfPoseEstimator etc. while running the command "python run_webcam.py --model=mobilenet_thin --resize=432x368 --camera=0" . I have build the pafprocess c++ files swig successfully, but also I'm getting the errors

  #43853 commented on Oct 8, 2020 • 2 new comments

• Possibility to compile python code when @tf.function is used in TensorFlow 2.x

  #43800 commented on Oct 8, 2020 • 2 new comments

• NotFoundError: No algorithm worked! when using Conv2D

  #43174 commented on Oct 9, 2020 • 2 new comments

• Keras: Regularizer not saved for Lambda layers

  #39900 commented on Oct 9, 2020 • 2 new comments

• Unexpected warning during GeneratorDataset iterator finalization

  #24570 commented on Oct 9, 2020 • 2 new comments

• TensorFlow hangs forever in multinode training with NCCL and certain model (with SyncBatchNorm layers)

  #42111 commented on Oct 10, 2020 • 2 new comments

• TFLite Failed to build gpu delegate on Linux

  #38336 commented on Oct 10, 2020 • 2 new comments

• ModelCheckpoint behavior doesn't change when changing save_weights_only from True to False for custom tf.keras.Model

  #42741 commented on Oct 10, 2020 • 2 new comments

• No gradient defined for operation RaggedTensorFromVariant / or no gradients at all

  #43107 commented on Oct 12, 2020 • 2 new comments

• TensorFlow 2.3.0 installation fails as protoc could not be found

  #41856 commented on Oct 12, 2020 • 2 new comments

• Getting a “RuntimeError: Expected bias tensor to be a vector” while trying to do full integer quantization on a keras model

  #43517 commented on Oct 12, 2020 • 2 new comments

• TFLiteCCoreML Delegate Runtime Crash (Espresso::ANERuntimeEngine overflow error)

  #43473 commented on Oct 13, 2020 • 2 new comments

• Decoding error during vocabulary loading from TextVectorization layer

  #43559 commented on Oct 13, 2020 • 2 new comments

• transpose() can be very slow on CPU

  #27383 commented on Oct 13, 2020 • 2 new comments

• Missing TRANSPOSE Op Kernel

  #43472 commented on Oct 13, 2020 • 2 new comments

• tf_record issue while following tutorial

  #43835 commented on Oct 13, 2020 • 2 new comments

• Unable to load graph from protobuf file on s390x arch (Big Endian)

  #41652 commented on Oct 13, 2020 • 2 new comments

• UnboundLocalError: local variable 'logs' referenced before assignment on training with little data

  #38064 commented on Oct 14, 2020 • 2 new comments

• TF2.0, how to include feature processing code in saved model.

  #43373 commented on Oct 14, 2020 • 2 new comments

• tf.keras.layers.LSTM + tf.function fails to compute jacobian with pfor on GPU

  #41042 commented on Oct 14, 2020 • 2 new comments

• RPC retries do not work on Istio

  #43838 commented on Oct 14, 2020 • 2 new comments

• [TF-TRT] Fix TopK Combined NMS

  #40062 commented on Oct 14, 2020 • 2 new comments

• Run entire epoch as compiled tf.function in model.fit()

  #41742 commented on Oct 14, 2020 • 2 new comments

• Noisy Networks for Exploration

  #42069 commented on Oct 13, 2020 • 2 new comments

• Added thickness parameter for drawing bounding boxes

  #42124 commented on Oct 14, 2020 • 2 new comments

• [INTEL MKL] Remove nGraph build support

  #42870 commented on Oct 14, 2020 • 2 new comments

• Allow dict input_spec for keras.Model

  #42969 commented on Oct 14, 2020 • 2 new comments

• Wrong Copyright/license info in core/lib/io files, originated in LevelDB

  #43018 commented on Oct 9, 2020 • 2 new comments

• Modifying get_optimizer_experimental_options

  #43342 commented on Oct 14, 2020 • 2 new comments

• add support for mips64 platform

  #43415 commented on Oct 12, 2020 • 2 new comments

• [INTEL MKL] Removed unnecessary omp_set_num() calls

  #43795 commented on Oct 12, 2020 • 2 new comments

• [INTEL MKL] MKL DNN 0.x code cleanup - Matmul op

  #43804 commented on Oct 14, 2020 • 2 new comments

• TFLu Apollo3 Maintenance

  #43831 commented on Oct 13, 2020 • 2 new comments

• Memory leaks when using tf.strings.split in map_func for tf.data.Dataset.map with eager execution.

  #35152 commented on Oct 8, 2020 • 1 new comment

• CMake issues when building TFLite

  #43646 commented on Oct 8, 2020 • 1 new comment

• Tensorflow C API: SessionRun batch size (how to properly set)

  #22869 commented on Oct 8, 2020 • 1 new comment

• Sparse input name missing in exported model

  #42018 commented on Oct 8, 2020 • 1 new comment

• Roadmap for publishing TFLite runtime on PyPi?

  #38642 commented on Oct 8, 2020 • 1 new comment

• cuda_configure.bzl: wrong cuda_config.cudnn_version cause wrong cudnn_headers list

  #43799 commented on Oct 9, 2020 • 1 new comment

• model.fit() is 10% slower over custom training loops when using TPUs

  #40453 commented on Oct 9, 2020 • 1 new comment

• In ModelCheckpoint, filepath is not accepting batch as formatting parameter.

  #38668 commented on Oct 9, 2020 • 1 new comment

• Cannot import tensorflow

  #43747 commented on Oct 9, 2020 • 1 new comment

• Example of inferencing a Tensorflow lite model with parsing_serving_input_receiver_fn using C++ API

  #43607 commented on Oct 9, 2020 • 1 new comment

• tf.py_function in tf.data.Dataset pipeline doesn't work with TPUEstimator

  #30818 commented on Oct 9, 2020 • 1 new comment

• Multiprocessing TypeError: can't pickle _thread.RLock objects, using tf.keras.model

  #42931 commented on Oct 9, 2020 • 1 new comment

• Float16 quantized DistilBERT model performs poorly

  #43719 commented on Oct 9, 2020 • 1 new comment

• Keras 'Random' preprocessing layers do not work on cloud TPU

  #42454 commented on Oct 9, 2020 • 1 new comment

• Large number is multiplied with the embedding output

  #42968 commented on Oct 9, 2020 • 1 new comment

• person_detection benchmarks do not build for STM32F4

  #43743 commented on Oct 9, 2020 • 1 new comment

• RNN document is not clear

  #34539 commented on Oct 9, 2020 • 1 new comment

• Non-OK-status: GpuLaunchKernel( SwapDimension1And2InTensor3UsingTiles<T, NumThreads, TileLongSide, TileShortSide>, ...) Internal: invalid configuration argument when using the tf.keras MaxPooling3D layer

  #36310 commented on Oct 9, 2020 • 1 new comment

• Error in the tpu.ipynb notebook

  #40339 commented on Oct 10, 2020 • 1 new comment

• Custom Implementation for FusedBatchNormV3

  #38113 commented on Oct 10, 2020 • 1 new comment

• Failed to convert SparseTensor to Tensor

  #42916 commented on Oct 10, 2020 • 1 new comment

• AttributeError: '_TfDeviceCaptureOp' object has no attribute 'node_def'

  #43323 commented on Oct 10, 2020 • 1 new comment

• TF 2.3 Saved model | Experimental features not working properly

  #42613 commented on Oct 10, 2020 • 1 new comment

• Hi, am trying to visualize the loss function using tensorboard

  #43756 commented on Oct 10, 2020 • 1 new comment

• You tried to call `count_params` on z_input, but the layer isn't built. You can build it manually via: `z_input.build(batch_input_shape

  #35353 commented on Oct 11, 2020 • 1 new comment

• Custom Keras model and layers demonstrate different behavior (nonzero vs zero grads) on CPU vs GPU

  #39610 commented on Oct 11, 2020 • 1 new comment

• Build windows x86 c++ library failed

  #41765 commented on Oct 11, 2020 • 1 new comment

• InternalError : failed to call ThenRNNBackward... on version 2.3 but not 2.1

  #42412 commented on Oct 11, 2020 • 1 new comment

• tf.keras.layers.experimental.preprocessing.CategoryEncoding: is it possible to apply it to a nD tensor?

  #43667 commented on Oct 11, 2020 • 1 new comment

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

  #43661 commented on Oct 11, 2020 • 1 new comment

• Negate Kernel for int8 is yet to be implemented

  #43425 commented on Oct 12, 2020 • 1 new comment

• Undefined symbol in debug mode

  #39521 commented on Oct 12, 2020 • 1 new comment

• null pointer dereference Error in TF2.3.0 with runforMultipleInputOutput

  #43657 commented on Oct 12, 2020 • 1 new comment

• Add metadata in tensorflow lite model

  #43775 commented on Oct 12, 2020 • 1 new comment

• Win10+CUDA11.1+cudnn8.0.4+rtx3090

  #43751 commented on Oct 12, 2020 • 1 new comment

• Cannot convert a Tensor of dtype resource to a NumPy array

  #42673 commented on Oct 12, 2020 • 1 new comment

• NASNetLarge Keras Model - accuracy is very very low

  #43730 commented on Oct 12, 2020 • 1 new comment

• ValueError: This model has not yet been built. Build the model first by calling `build()` or calling `fit()` with some data, or specify an `input_shape` argument in the first layer(s) for automatic build

  #43709 commented on Oct 12, 2020 • 1 new comment

• Convert saved model - issue generated shapes

  #42153 commented on Oct 12, 2020 • 1 new comment

• How it is done convolution calculation order in conv2d?

  #43724 commented on Oct 12, 2020 • 1 new comment

• LookupError: gradient registry has no entry for: TensorScatterMax

  #43765 commented on Oct 12, 2020 • 1 new comment

• Serial printing breaks after updating Mbed OS

  #43821 commented on Oct 12, 2020 • 1 new comment

• Keras model pickle-able but tf.keras model not pickle-able

  #34697 commented on Oct 12, 2020 • 1 new comment

• Evaluation produces OSError: [Errno 24] Too many open files

  #43738 commented on Oct 12, 2020 • 1 new comment

• Keras batch norm performance issue with parameter server strategy

  #43638 commented on Oct 12, 2020 • 1 new comment

• Accessing validation data within a custom callback ??

  #32981 commented on Oct 12, 2020 • 1 new comment

• overlap backward and optimization in TensorFlow 2.0

  #33274 commented on Oct 12, 2020 • 1 new comment

• ConnectionResetByPeer using MultiWorkerMirroredStrategy with Native Keras BERT Model

  #42498 commented on Oct 12, 2020 • 1 new comment

• AssertionError on using model.fit() in tf.distribute.MirroredStrategy

  #37485 commented on Oct 12, 2020 • 1 new comment

• scaling issue by using multiworkerstrategy for CPU

  #42888 commented on Oct 12, 2020 • 1 new comment

• Lambda layer does not compute masked values properly

  #40085 commented on Oct 12, 2020 • 1 new comment

• Many context switches / Many threads even if threading is limited

  #29968 commented on Oct 12, 2020 • 1 new comment

• Keras Backend ones_like with Lambda is not serializable

  #43745 commented on Oct 12, 2020 • 1 new comment

• creating segmentation

  #43760 commented on Oct 12, 2020 • 1 new comment

• Named dictionary inputs and outputs for tf.keras.Model

  #34114 commented on Oct 12, 2020 • 1 new comment

• How about we stop deprecating everything every other release....

  #43303 commented on Oct 12, 2020 • 1 new comment

• saved_model_cli fails when --input_examples option contains string feature

  #27662 commented on Oct 13, 2020 • 1 new comment

• Support "fetch_skip_sync=false" in Callable

  #33187 commented on Oct 13, 2020 • 1 new comment

• Feature request: GPU ops for tf.unique; tf.where; and tf.dynamic_partition

  #5965 commented on Oct 13, 2020 • 1 new comment

• post quant keras model with UpSample2D ，after post-quantized get a false output shape

  #43595 commented on Oct 13, 2020 • 1 new comment

• TensorFlow Lite Error: tensorflow/lite/kernels/concatenation.cc:52 axis >= 0 was not true.

  #37733 commented on Oct 13, 2020 • 1 new comment

• Unable to train the model

  #43736 commented on Oct 13, 2020 • 1 new comment

• Slow training with tfRecords and tf.functions

  #43585 commented on Oct 13, 2020 • 1 new comment

• RTX 3090 Cuda 10.1 Kernal image error

  #43718 commented on Oct 13, 2020 • 1 new comment

• tf.summary.create_file_writer does not write any event file and tensorboard show nothing

  #43584 commented on Oct 13, 2020 • 1 new comment

• My code is much slower in TF2 than TF1

  #43677 commented on Oct 13, 2020 • 1 new comment

• While Tensorflow usually uses GatherV2, TFLite only supports Gather

  #42180 commented on Oct 13, 2020 • 1 new comment

• Dense variable constraints are not allowed with sparse gradients

  #33755 commented on Oct 13, 2020 • 1 new comment

• Threading + tf.keras.layers.Input produces TypeError: 'NoneType' object is not iterable

  #35813 commented on Oct 14, 2020 • 1 new comment

• Gradient Compute Error in Embedding Layers

  #38130 commented on Oct 14, 2020 • 1 new comment

• Custom optimizer that behaves differently depending on the shapes of weights

  #39298 commented on Oct 14, 2020 • 1 new comment

• tensorflow cannot use feature columns to keras model

  #39818 commented on Oct 14, 2020 • 1 new comment

• 'fused' argument of BatchNormalization is not saved in the model

  #40224 commented on Oct 14, 2020 • 1 new comment

• mAP drops crazily when replacing Conv with Depthwise_Conv

  #43256 commented on Oct 14, 2020 • 1 new comment

• Tensorflow 2.2 takes much more time than 2.1/2.0 to start training with "keras.fit"

  #40036 commented on Oct 14, 2020 • 1 new comment

• TensorFlow Lite causes segementation faults in C++ when creating threads if fully statically linked

  #42057 commented on Oct 14, 2020 • 1 new comment

• Openpose installation issue

  #43843 commented on Oct 14, 2020 • 1 new comment

• Equivalent of Torch EmbeddingBag

  #32675 commented on Oct 14, 2020 • 1 new comment

• TF hangs on HTC Cluster

  #43656 commented on Oct 14, 2020 • 1 new comment

• TFLite ConverterError: Node has inputs from different frames

  #43849 commented on Oct 14, 2020 • 1 new comment

• "No gradients provided" with GradientDescentOptimizer if not using float variable

  #43845 commented on Oct 14, 2020 • 1 new comment

• Ability to calculate projected memory usage for a given model

  #36327 commented on Oct 14, 2020 • 1 new comment

• Defined layers but received error: __init__() missing 1 required positional argument: 'layer'

  #43771 commented on Oct 14, 2020 • 1 new comment

• Close (and reopen new) summary files at regular intervals

  #38203 commented on Oct 13, 2020 • 1 new comment

• [TFLite] Fully quantize the uint8 concatenation kernel to make it pure integer

  #38704 commented on Oct 13, 2020 • 1 new comment

• Add missing double overloads for GetAttr, GetNodeAttr, and TryGetNodeAttr

  #39452 commented on Oct 14, 2020 • 1 new comment

• Change snappy_buffers_test record size for s390x

  #40198 commented on Oct 14, 2020 • 1 new comment

• TF-TRT Improve reshape op converter

  #40545 commented on Oct 9, 2020 • 1 new comment

• keras preprocessing layer random generator fix

  #41143 commented on Oct 8, 2020 • 1 new comment

• Improve ROCm's sqrt and rsqrt for std::complex.

  #41303 commented on Oct 14, 2020 • 1 new comment

• Fix invalid fusion of Matmul and Mul

  #41790 commented on Oct 13, 2020 • 1 new comment

• Fix segmentation fault in tf.image.crop_and_resize when boxes is inf or nan

  #42143 commented on Oct 14, 2020 • 1 new comment

• [TFLite 16x8] Fixes for TANH and LOGISTIC

  #42671 commented on Oct 9, 2020 • 1 new comment

• TFLu: Fix Ethos-U build issue

  #42823 commented on Oct 13, 2020 • 1 new comment

• C++ canonicalize_function_inputs WIP

  #43053 commented on Oct 14, 2020 • 1 new comment

• [ROCm] Disabling failing unit tests

  #43159 commented on Oct 14, 2020 • 1 new comment

• [MLIR] Add involution based folding to TF

  #43428 commented on Oct 14, 2020 • 1 new comment

• Add transform to fuse activations into TFL pooling ops

  #43482 commented on Oct 14, 2020 • 1 new comment

• [tflite] fix missing labels.txt in model maker tutorial

  #43488 commented on Oct 14, 2020 • 1 new comment

• TFLu: Allocate scratch tensors even though they are not in subgraph

  #43538 commented on Oct 13, 2020 • 1 new comment

• TFLu: Port TFL detection postprocess operator

  #43566 commented on Oct 14, 2020 • 1 new comment

• Solve IndexError: list index out of range #43561

  #43570 commented on Oct 10, 2020 • 1 new comment

• WorkerTrainingState Test and Bugfix

  #43592 commented on Oct 14, 2020 • 1 new comment

• Adding README.md to TFLite evaluation/tasks

  #43675 commented on Oct 12, 2020 • 1 new comment

• Fix bug: "Unsupported activation type for quantize-activation: 0"

  #43731 commented on Oct 14, 2020 • 1 new comment

• Cannot save to SavedModel when custom layers accepts args and kwargs

  #38384 commented on Oct 7, 2020 • 0 new comments

• Saving and loading Keras model with RNN layer that uses both multiple input and constants will result in ValueError when using model.predict()

  #43773 commented on Oct 7, 2020 • 0 new comments

• When dose Tensorflow add elasticity function in worker？

  #41589 commented on Oct 8, 2020 • 0 new comments

• Segmentation fault (core dumped)

  #43753 commented on Oct 8, 2020 • 0 new comments

• Correct way to cast a RaggeredTensor to list of strings?

  #43805 commented on Oct 8, 2020 • 0 new comments

• Model conversion fails with cryptic error when feeding sample during post-training quantization

  #43239 commented on Oct 8, 2020 • 0 new comments

• Didn't find op for builtin opcode 'CONV_2D' version '5'

  #43232 commented on Oct 8, 2020 • 0 new comments

• Need use cases to guide tensorflow installation

  #43519 commented on Oct 8, 2020 • 0 new comments

• Access to GPU memory usage

  #43846 commented on Oct 8, 2020 • 0 new comments

• Cannot convert Mobile Zoo tf2 models to tflite

  #43495 commented on Oct 8, 2020 • 0 new comments

• Unable to access files on S3 with tf.io.gfile.GFile

  #38054 commented on Oct 8, 2020 • 0 new comments

• TFLite android image classification demo failed on samsung note10 whose soc is snapdragon 855

  #43663 commented on Oct 9, 2020 • 0 new comments

• InvalidArgumentError with `Conv1D` which has the same `dilation_rate` as the batch size, using `tf.keras.Model.fit`, on TPU

  #43690 commented on Oct 9, 2020 • 0 new comments

• tf.gather Converting sparse IndexedSlices warning

  #38168 commented on Oct 9, 2020 • 0 new comments

• Batch Renormalization via Keras Layer not working with TF 2+ and tf.function

  #43732 commented on Oct 9, 2020 • 0 new comments

• Google Collab error for TPU - UnavailableError: {{function_node __inference_train_function_99378}} failed to connect to all addresses

  #43790 commented on Oct 9, 2020 • 0 new comments

• TypeError in "Bidirectional.compute_output_shape"

  #43816 commented on Oct 9, 2020 • 0 new comments

• tensorflow.vectorized_map cannot recive loop variable

  #43822 commented on Oct 9, 2020 • 0 new comments

• Keras `model_from_json` ignores distribution strategy

  #38396 commented on Oct 9, 2020 • 0 new comments

• can't compile tensorflow from source

  #43782 commented on Oct 9, 2020 • 0 new comments

• divide_no_nan logic

  #43622 commented on Oct 9, 2020 • 0 new comments

• Possible GPU memory leak at tf.vectorized_map

  #43593 commented on Oct 9, 2020 • 0 new comments

• import error -> tensorflow

  #43431 commented on Oct 9, 2020 • 0 new comments

• tf.nn.fractional_max_pool aborts when pooling_ratio is negative

  #42278 commented on Oct 9, 2020 • 0 new comments

• TensorBoard callback doesn't update step properly

  #43568 commented on Oct 9, 2020 • 0 new comments

• Single threaded TensorFlow

  #43671 commented on Oct 9, 2020 • 0 new comments

• XLA Compilation does not work with Embeddings Layer

  #43687 commented on Oct 9, 2020 • 0 new comments

• XLA Compilation Bug

  #43723 commented on Oct 9, 2020 • 0 new comments

• No library found under: /usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudart.so.11.1

  #43689 commented on Oct 9, 2020 • 0 new comments

• IteratorGetNext: unsupported op: No registered 'IteratorGetNext' OpKernel for XLA_GPU_JIT devices compatible with node {{node IteratorGetNext}} . Registered: device='XLA_GPU'

  #43665 commented on Oct 9, 2020 • 0 new comments

• KeyError of identity node, when using Skip Connection and RandomDrop together

  #36712 commented on Oct 9, 2020 • 0 new comments

• tf.keras.applications.EfficientNetB- not running as expected

  #42506 commented on Oct 10, 2020 • 0 new comments

• Tensorflow lite speech sample is not detecting silence (wrong detections)

  #43340 commented on Oct 10, 2020 • 0 new comments

• Unexpected output shape when trying to convert to Frozen Graph using convert_variables_to_constants_v2

  #43839 commented on Oct 10, 2020 • 0 new comments

• Undefined symbols for architecture arm64 when loading TensorFlowLiteSelectTfOps on iOS device

  #41948 commented on Oct 11, 2020 • 0 new comments

• CUDA 10.1

  #43733 commented on Oct 12, 2020 • 0 new comments

• Problem with custom grad with multiple external variables

  #43535 commented on Oct 13, 2020 • 0 new comments

• Custom loss function is not working

  #43650 commented on Oct 14, 2020 • 0 new comments

• Add Usage Example to Keras Nadam Optimizer

  #35984 commented on Oct 13, 2020 • 0 new comments

• Add and populate leaf_id in boosted_trees.DebugOutput

  #37455 commented on Oct 14, 2020 • 0 new comments

• Fix only pass RunOptions to keras will trigger core

  #38562 commented on Oct 14, 2020 • 0 new comments

• Add __reduce_ex__ to Keras Model to enable copy.deepcopy and pickle

  #39609 commented on Oct 9, 2020 • 0 new comments

• Const-qulify PyUFuncGenericFunction's arguments

  #40728 commented on Oct 9, 2020 • 0 new comments

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

  #41441 commented on Oct 14, 2020 • 0 new comments

• [TFLite 16x8] ADD/SUB operators: fixes + tests for versioning

  #42373 commented on Oct 9, 2020 • 0 new comments

• Ported TFLM Micro Speech example to CEVA-DSP SP500

  #42469 commented on Oct 9, 2020 • 0 new comments

• CallCounter C++ version

  #42540 commented on Oct 9, 2020 • 0 new comments

• [Intel MKL] unit test for bfloat16 on matmul.

  #42949 commented on Oct 7, 2020 • 0 new comments

• Overview sanilty check fails on pylint upgrade

  #43040 commented on Oct 14, 2020 • 0 new comments

• Revert renaming of tools to exec_tools

  #43156 commented on Oct 14, 2020 • 0 new comments

• Uses of the new rocm skip test decorator.

  #43161 commented on Oct 14, 2020 • 0 new comments

• Integrate cuBLASLt API into backend

  #43237 commented on Oct 9, 2020 • 0 new comments

• Update the CMSIS-NN glue code to support int16 for softmax.

  #43545 commented on Oct 14, 2020 • 0 new comments

• enlarge the format data width

  #43587 commented on Oct 14, 2020 • 0 new comments

• [TFLite] Add int16x8 support for non-fused RELU and RELU6 operators

  #43617 commented on Oct 14, 2020 • 0 new comments

• [TFLite] Add int16x8 support for REDUCE_MIN and REDUCE_MAX operators

  #43618 commented on Oct 14, 2020 • 0 new comments

• [ROCm] Porting changes to support newer ROCm versions to r2.1

  #43633 commented on Oct 13, 2020 • 0 new comments

• [Draft] Adding Multi device iterator Serialization and De-serialization

  #43695 commented on Oct 8, 2020 • 0 new comments

• Fixed inputs not being cast to tuple

  #43774 commented on Oct 14, 2020 • 0 new comments

• [INTEL MKL] MKL DNN 0.x code cleanup - Relu ops

  #43780 commented on Oct 9, 2020 • 0 new comments

• fix grappler/costs:op_performance_data dependency

  #43823 commented on Oct 13, 2020 • 0 new comments

• [coreml delegate] fix usage of coremltools header

  #43837 commented on Oct 13, 2020 • 0 new comments