tensorflow / tensorflow

1 Release published by 1 person

• v2.4.0 TensorFlow 2.4.0

  published Dec 14, 2020

61 Pull requests merged by 29 people

• [ROCm] Updates required to add ROCm support for JAX

  #45583 merged Dec 16, 2020

• Fix sanity build

  #45729 merged Dec 16, 2020

• Fix sanity build

  #45728 merged Dec 16, 2020

• [TFTRT - Dynamic Shape Phase 3] Add Dynamic Shape Testing for ConvertClipByValue

  #45589 merged Dec 16, 2020

• Add check to ensure that files in tensorflow/lite/micro have proper license and formatting

  #45674 merged Dec 15, 2020

• Add upper bound to `h5py`.

  #45712 merged Dec 15, 2020

• Add upper bound to `h5py`.

  #45711 merged Dec 15, 2020

• Add upper bound to `h5py`.

  #45710 merged Dec 15, 2020

• Fix sanity build, reorder tags to satisfy buildifier.

  #45724 merged Dec 15, 2020

• Disable a few tests.

  #45726 merged Dec 15, 2020

• Disable a few tests.

  #45725 merged Dec 15, 2020

• Disable a few tests.

  #45723 merged Dec 15, 2020

• Disable a few tests.

  #45722 merged Dec 15, 2020

• Disable a few tests.

  #45721 merged Dec 15, 2020

• micro: fix build warning due to example person_detection

  #45666 merged Dec 15, 2020

• Refactor SparseApply[Proximal]Adagrad into classes

  #44919 merged Dec 15, 2020

• fix asan check in c++ gradients

  #45654 merged Dec 15, 2020

• Fix bazel build with -DTF_LITE_STATIC_MEMORY.

  #45679 merged Dec 15, 2020

• Fixed incorrect ptxas usage affecting performance when using CUDA

  #44486 merged Dec 15, 2020

• README: update Linaro CI links

  #45595 merged Dec 15, 2020

• update estimator version after estimator final release

  #45677 merged Dec 15, 2020

• Add missing arch-specific dependencies to tf_to_kernel

  #45265 merged Dec 15, 2020

• Download pigweed as part of the TensorflowLite Micro makefile.

  #45672 merged Dec 15, 2020

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

  #44022 merged Dec 15, 2020

• Manually fix licenses for the files to ensure consistent formatting.

  #45673 merged Dec 15, 2020

• Refactor quantize (reference and xtensa) for a complete fallback.

  #45618 merged Dec 14, 2020

• Making label smoothing documentation more helpful

  #45215 merged Dec 14, 2020

• Update local.py

  #45400 merged Dec 14, 2020

• [Intel MKL] Add MatMul+BiasAdd+Add fusion

  #45358 merged Dec 14, 2020

• Update release notes for TensorFlow 2.4.0

  #44220 merged Dec 12, 2020

• [TFTRT] Rename ParameterizedOpConverterTest classes with more descriptive names

  #45587 merged Dec 12, 2020

• Cleanup `tensorflow/c/experimental/gradients` Part 1

  #45547 merged Dec 11, 2020

• Keras SavedModel: Ignore custom metrics failure when compile=False

  #45278 merged Dec 11, 2020

• Try to switch to h5py 3.1.0

  #45487 merged Dec 11, 2020

• Only use site.USER_SITE if ENABLE_USER_SITE is set

  #45399 merged Dec 11, 2020

• [Intel MKL] Improve CNMS performance by removing unnecessary allocation

  #44923 merged Dec 11, 2020

• [tflite] map leaky relu to prelu to use NNAPI

  #45296 merged Dec 11, 2020

• Fix unique shape inference

  #45280 merged Dec 11, 2020

• Changed TOSA common legalizations to use llvm::Optional<>

  #45496 merged Dec 11, 2020

• Extract a function for parsing operator DIV

  #45440 merged Dec 10, 2020

• [TFLite] Fix potential overflow in 64-bit MultiplyByQuantizedMultiplier function

  #42134 merged Dec 10, 2020

• ellaborated docs on determinism in dataset ops

  #45495 merged Dec 10, 2020

• Dispatch reverse_sequence

  #45534 merged Dec 10, 2020

• Run apt-get update before installing jupyter dependencies

  #45555 merged Dec 10, 2020

• Update version numbers for TensorFlow 2.4.0

  #45559 merged Dec 10, 2020

• Dockerfiles: Pin pip version and fix apt-get

  #45562 merged Dec 10, 2020

• Add GPU kernel for SparseApplyFtrl

  #44540 merged Dec 10, 2020

• Dockerfiles: Pin pip version and fix apt-get

  #45561 merged Dec 9, 2020

• Dockerfiles: Pin pip version and fix apt-get

  #45560 merged Dec 9, 2020

• tflite: Add casts to std::min/std::max when comparing mis-matched types

  #45486 merged Dec 9, 2020

• Update PCRE library from 8.42 to 8.44

  #45553 merged Dec 9, 2020

• Build for TARGET_ARCH=fusion_f1 via reference implementation fallback.

  #45464 merged Dec 9, 2020

• Prevent unitialized memory access in `GraphConstructor::MakeEdge`

  #45503 merged Dec 9, 2020

• Prevent CHECK-fail in LSTM/GRU with zero-length input.

  #45502 merged Dec 9, 2020

• Completely rewrite `GetMatchingPaths`.

  #45501 merged Dec 9, 2020

• Mark `MemmappedTensorAllocator` as returning opaque handle.

  #45500 merged Dec 9, 2020

• Validate that `DataFormat*` attributes form a permutation.

  #45499 merged Dec 9, 2020

• Default initialize fixed point Eigen types.

  #45498 merged Dec 9, 2020

• Downgrade Docker container Pip to <20.3

  #45491 merged Dec 9, 2020

• [ROCm] Re-enabling unit-tests that are now passing on ROCm platform

  #45386 merged Dec 9, 2020

• TFLu: Fix bug in PPD op

  #45474 merged Dec 9, 2020

50 Pull requests opened by 35 people

• Fix Windows platform file stat length out of range error

  #45542 opened Dec 9, 2020

• remove $ from readme for easy compatibility with Windows

  #45544 opened Dec 9, 2020

• Revert "Polish TensorRT static linking a little."

  #45545 opened Dec 9, 2020

• Update PCRE library from 8.42 to 8.44

  #45548 opened Dec 9, 2020

• Update PCRE library from 8.42 to 8.44

  #45549 opened Dec 9, 2020

• Update PCRE library from 8.42 to 8.44

  #45550 opened Dec 9, 2020

• Update PCRE library from 8.42 to 8.44

  #45551 opened Dec 9, 2020

• Update PCRE library from 8.42 to 8.44

  #45552 opened Dec 9, 2020

• [INTEL MKL] Enable DNNL cpu dispatch control

  #45554 opened Dec 9, 2020

• Adding Configurable Filesystem API part 1

  #45563 opened Dec 10, 2020

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

  #45564 opened Dec 10, 2020

• Update README.md

  #45567 opened Dec 10, 2020

• implement GetCurrentClockCycle() for s390x

  #45576 opened Dec 10, 2020

• Add TF-TRT optimization profiles for testing

  #45588 opened Dec 10, 2020

• [Intel MKL] Fix memory leak in MklAddN

  #45593 opened Dec 11, 2020

• TF-TRT ConvertPad in dynamic shape mode

  #45597 opened Dec 11, 2020

• TF-TRT Test ConvertQuantize in dynamic shape mode

  #45599 opened Dec 11, 2020

• [XLA] [Docs] Fix typo in operation semantics

  #45600 opened Dec 11, 2020

• TFLM Added optimized fully connected (float32 and int8) for CEVA-BX1

  #45606 opened Dec 11, 2020

• Fix convert_variables_to_constants_v2 for scalar valued captures

  #45610 opened Dec 11, 2020

• Fix the issue of tf.image.central_crop when central_fraction is a tensor

  #45613 opened Dec 11, 2020

• Only use site.USER_SITE if site.ENABLE_USER_SITE is set

  #45614 opened Dec 11, 2020

• Build Xtensa kernels for Vision P6

  #45619 opened Dec 12, 2020

• [TFTRT] Disable non TRT optimization fix

  #45620 opened Dec 12, 2020

• Add keras.metrics.HarmonicMean

  #45638 opened Dec 13, 2020

• __call__ will call build

  #45639 opened Dec 13, 2020

• activation, kernel, bias are all attributes in Dense

  #45640 opened Dec 13, 2020

• TFLM: Adding TANH op optimized for CEVA-BX1

  #45641 opened Dec 13, 2020

• Add classifier_activation param to applications

  #45643 opened Dec 14, 2020

• micro: remove lite-specific code from copy of FILL

  #45646 opened Dec 14, 2020

• micro: port operator FILL from lite

  #45647 opened Dec 14, 2020

• Use a raw string for docstring with TeX

  #45655 opened Dec 14, 2020

• Prevent crash of tensorflow if shape is too large for tf.sparse.reorder

  #45675 opened Dec 14, 2020

• fix `modular_filesystem` call itself

  #45682 opened Dec 15, 2020

• Remove requirement that first dimensions of inputs are equal when using train_on_batch and test_on_batch

  #45689 opened Dec 15, 2020

• MeanIoU: Support y_pred passed as a logits tensor

  #45690 opened Dec 15, 2020

• Add Layer Dense2d

  #45691 opened Dec 15, 2020

• Fix for TensorFlow Lite to parse tensor shape signature safely on big endian machines

  #45694 opened Dec 15, 2020

• Refactor flatbuffer_conversions.cc/h for BATCH_TO_SPACE_ND and SPACE_TO_BATCH_ND

  #45696 opened Dec 15, 2020

• [ROCm] Updating Stream Executor Blas interface (for ROCm) to call rocblas APIs…

  #45697 opened Dec 15, 2020

• Refactor reference implementation from lite/kernels/internal/reference/reference_ops.h into own headers

  #45699 opened Dec 15, 2020

• copy reference kernels from lite to micro

  #45704 opened Dec 15, 2020

• Add Cuda 11.0 custom-op Dockerfile

  #45714 opened Dec 15, 2020

• Add Cuda 11 custom-op Dockerfile

  #45715 opened Dec 15, 2020

• [Cherrypick:r2.3] Don't do fused average updates inside XLA context as it may create extra tf.cond which causes OOM on TPUs.

  #45718 opened Dec 15, 2020

• [Cherrypick:r2.2] SavedModel loading: Create temporary graphs for importing functions to avoid leaking memory

  #45719 opened Dec 15, 2020

• Some initial documentation for the TfLite Micro CI system.

  #45720 opened Dec 15, 2020

• Update tflite-converter-issue template

  #45732 opened Dec 16, 2020

• changed a sentence in readme to make it more easy to understand

  #45734 opened Dec 16, 2020

• Expose transform

  #45742 opened Dec 16, 2020

96 Issues closed by 58 people

• RNN Using TensorFlow 2.0

  #45628 closed Dec 16, 2020

• KeyError: "The name 'import/final_result' refers to an Operation not in the graph."

  #45303 closed Dec 16, 2020

• illegal hardware instruction python while import tensorflow2.2.0

  #45323 closed Dec 16, 2020

• what are the .pb and .h5 model file?

  #45644 closed Dec 16, 2020

• @Tuxius and others which stuck until this bug is fixed:

  #45737 closed Dec 16, 2020

• Unable to install CUDA on Ubuntu 18.04.5

  #44302 closed Dec 16, 2020

• ops defined inside tf.while_loop's cond/body or tf.cond's true_fn/false_fn functions ignore their enclosed tf.device if the tf.while_loop/tf.cond itself is inside a tf.device

  #44011 closed Dec 16, 2020

• Build tensorflow from source, issues about CUDA lib and header paths

  #45667 closed Dec 16, 2020

• ERROR on running frozen.pb at tf2.x (google.protobuf.message.DecodeError: Error parsing message )

  #45685 closed Dec 16, 2020

• micro: make build warnings cause false-positives while browsing compiler errors via editor

  #45665 closed Dec 15, 2020

• ""TypeError: 'NoneType' object is not iterable"" when instantiating keras.models.Model() or keras.models.Sequential()

  #45578 closed Dec 15, 2020

• Tflite Model gives same output for every input

  #45119 closed Dec 15, 2020

• Can't load Model with SparseTensor inputs if it calls a Sequential model internally

  #45652 closed Dec 15, 2020

• QAT conversion RuntimeError: Quantization not yet supported for op: 'DEQUANTIZE' issue with tf-nightly

  #42082 closed Dec 15, 2020

• BoostedTreesClassifier segmentation fault when evaluate or predict is called

  #45603 closed Dec 15, 2020

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

  #40036 closed Dec 15, 2020

• batch_gather_nd op <del>Size Op</del> missing for Tensorflow Lite

  #34101 closed Dec 15, 2020

• Cannot build libtensorflowlite_c.so -- problem with bazel build

  #33513 closed Dec 15, 2020

• Extending ctc_batch_cost to handle RaggedTensors

  #41536 closed Dec 15, 2020

• Use pretrained weight for another smaller model

  #45271 closed Dec 15, 2020

• Namerror when loading model

  #45272 closed Dec 15, 2020

• Models loaded from files do not perform as before saving

  #43230 closed Dec 15, 2020

• Is there a tutorial and guide implement with c ？

  #45393 closed Dec 15, 2020

• Missing dependency on PowerPCCodeGen leads to undefined references in tf_to_kernel

  #45104 closed Dec 15, 2020

• ERROR:tensorflow: Failed to close session after error.Other threads may hang.

  #44824 closed Dec 14, 2020

• Plans to release 2.4.0 as a stable release?

  #44616 closed Dec 14, 2020

• Which version of Protobuf do I need to compile the TensorFlow2.0 source code ？

  #45083 closed Dec 14, 2020

• RTX3070 CUDA 11.1 CuDNN 8.0.5 not GPU available with tensorflow 2.5 (nightly version)

  #45653 closed Dec 14, 2020

• tensorflow 2.4 depends on tensorflow-estimator 2.4 which was not released in tandem

  #45669 closed Dec 14, 2020

• Model optimization using quantization on desktop PC?

  #45413 closed Dec 14, 2020

• addv2

  #45447 closed Dec 14, 2020

• Is there any support anywhere for Tensorflow on Apple Silicon?

  #45631 closed Dec 14, 2020

• torch.nn.functional.grid_smaple tensorflow2.0 how to recurrence?

  #45270 closed Dec 14, 2020

• DLL load failed while importing _pywrap_tensorflow_internal

  #45273 closed Dec 14, 2020

• Not working GeForce GTX 1070 on Tensorflow 2.3.1

  #45062 closed Dec 14, 2020

• GradientTape.batch_jacobian returns the wrong type in case of constant functions

  #43043 closed Dec 14, 2020

• TFLite, 2.2.0, accuracy drops significantly when tf.lite.Optimize.DEFAULT option is used

  #40000 closed Dec 14, 2020

• Layers with non-zero gradients are not updated with optimizer.apply_gradients()

  #45582 closed Dec 14, 2020

• Deprecated get_losses_for function in migration guide

  #43341 closed Dec 14, 2020

• tf.SparseTensorSpec in input_signature of tf.function does not take effect

  #45565 closed Dec 14, 2020

• Training a tf.keras.Model, with a tf.data.Dataset mapped with a randomly selected action from a large list of actions (callables provided by the user) takes a lot of time to start, if it ever does.

  #45162 closed Dec 14, 2020

• Slower prediction of tflite model then expected

  #45226 closed Dec 14, 2020

• No OpKernel was registered to support Op 'BroadcastTo' used by {{node gradients/Mean_grad/BroadcastTo}}with these attrs: [Tidx=DT_INT32, T=DT_FLOAT]

  #45472 closed Dec 14, 2020

• keras load_model Could not find matching function when compute_mask is used

  #45602 closed Dec 14, 2020

• A hierarchical model for device placement.

  #45225 closed Dec 14, 2020

• Invalid argument: During Variant Host->Device Copy: non-DMA-copy attempted of tensor type: string

  #45395 closed Dec 14, 2020

• NaN predictions/gradients/losses in basic GAN training

  #45623 closed Dec 13, 2020

• Object Detection doesn't work

  #45586 closed Dec 13, 2020

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

  #40224 closed Dec 13, 2020

• tokenizer = tfds.features.text.Tokenizer(),error is has no attribute 'text'.

  #45217 closed Dec 12, 2020

• Tensorflow float32 conversion bug from python to tensorflow constants

  #45596 closed Dec 12, 2020

• UnicodeDecodeError: 'utf-8' codec can't decode byte 0xXX in position YY : invalid continuation byte

  #45617 closed Dec 12, 2020

• Allow device groups / CPU in MirroredStrategy

  #33645 closed Dec 12, 2020

• tfjs-models/face-landmarks-detection/demo/ TypeError: Fail to fetch

  #45591 closed Dec 12, 2020

• tf2.3 keras.models.load_model setting compile=False fails to load saved_model but tf2.0 works.

  #43478 closed Dec 11, 2020

• Tensor data is stored in BE format on s390x in flatbuffer model in a certain scenario

  #45211 closed Dec 11, 2020

• h5py==3.0.0 causes issues with keras model loads in tensorflow 2.1.0

  #44467 closed Dec 11, 2020

• Multiple version of tensorflow that exist in system site packages and venv site packages

  #42978 closed Dec 11, 2020

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

  #42931 closed Dec 11, 2020

• Unexpected `snapshot` behaviour with `flat_map` in tf-nightly

  #44278 closed Dec 11, 2020

• cuDNN issue in TF

  #45423 closed Dec 11, 2020

• bad shape of index returned by UniqueWithCountsV2 in graph mode

  #44788 closed Dec 11, 2020

• SSD Mobilenet v2 model fail in PPD op

  #45473 closed Dec 11, 2020

• built wheel is not complete

  #45336 closed Dec 11, 2020

• How can i increase number of timeseries feature when building Encoder-Decoder model(seq2seq) using ConvLSTM2D

  #45206 closed Dec 11, 2020

• InvalidArgumentError: In[0] is not a matrix. Instead it has shape [] [Op:MatMul]

  #44359 closed Dec 11, 2020

• RMSprop fails if using mixed precision training and eager functions

  #45536 closed Dec 11, 2020

• deterministic `Dataset.map`s aren't

  #44491 closed Dec 10, 2020

• Keras: inconsistent behavior of "embedded" models

  #45437 closed Dec 10, 2020

• DLL error importing python tensorflow library in pywrap_tensorflow.py

  #44272 closed Dec 10, 2020

• Cannot load saved model when using multiple tf.keras.layers.experimental.preprocessing.StringLookup layers

  #45580 closed Dec 10, 2020

• tensorflow.python.framework.errors_impl.UnknownError: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above. [Op:Conv2D]

  #45458 closed Dec 10, 2020

• keras model to a single pb file or valid graph

  #45454 closed Dec 10, 2020

• Freeze / crash occurring when batch size is reduced

  #44459 closed Dec 10, 2020

• Could not load dynamic library 'libcudart.so.11.0'

  #44583 closed Dec 10, 2020

• Android app keep crashing when executing `interpreter.run(input, output)`

  #42850 closed Dec 10, 2020

• Migrating from TF1 to TF2, loss not going down using F-RCNN Resnet50

  #45192 closed Dec 10, 2020

• TensorFlow Lite Android nightly method "runForMultipleInputsOutputs" that loaded model using stft not work with converted by tf.lite.OpsSet.SELECT_TF_OPS

  #42723 closed Dec 10, 2020

• TensorFlow2 tutorials Actor-Critic method

  #43029 closed Dec 10, 2020

• Issue in RECURRENT NEURAL NETWORK

  #45411 closed Dec 10, 2020

• Beginner models

  #35608 closed Dec 10, 2020

• When installed by pip, dependencies' versions are not fixed.

  #45510 closed Dec 10, 2020

• Contributing to TFlite Micro

  #45350 closed Dec 10, 2020

• TFLite java.lang.IllegalArgumentException: Internal error: Failed to run on the given Interpreter on Android SDK 29

  #45540 closed Dec 10, 2020

• CONV_2d convert to DEPTHWISE_CONV when input depth=1

  #44532 closed Dec 9, 2020

• Conversion to TFLite fails with 2.4.0-rc1

  #44975 closed Dec 9, 2020

• Page is not visible in Desktops

  #45429 closed Dec 9, 2020

• [TFLite, Bug] Unexpected weights changes after conversion

  #33613 closed Dec 9, 2020

• Cannot ./configure Tensorflow on Ubuntu 18.04 LTS with CUDA 11.1

  #45463 closed Dec 9, 2020

• > It work for me.Thank you

  #45497 closed Dec 9, 2020

• Why do you want to ruin go's library?

  #45320 closed Dec 9, 2020

• tensorflow linear regression error blows up

  #45167 closed Dec 9, 2020

• Strange Out Of Memory error when trying example Colab

  #45468 closed Dec 9, 2020

• TF Lite LSH Projection returns different results on Big Endian and Little Endian Machines

  #44982 closed Dec 9, 2020

• `GetIntPtr` in `tensorflow/lite/tools/verifier.cc` do not get correct value on Big Endian machine

  #44852 closed Dec 9, 2020

• TypeError: 'NoneType' object is not subscriptable when training with tf.dataset

  #45533 closed Dec 9, 2020

95 Issues opened by 87 people

• Will tf.reshape() perform extra memcopy in tensorflow gpu backend?

  #45741 opened Dec 16, 2020

• tfLite armv7 build Issue

  #45740 opened Dec 16, 2020

• GET returned 404 Not Found

  #45739 opened Dec 16, 2020

• Why is the object_C library empty

  #45738 opened Dec 16, 2020

• model.predict(x,batch_size=500) for around half an hour report WebSocketClosedError()

  #45736 opened Dec 16, 2020

• No tensorflow lite 2.4.0 pre-built libraries

  #45735 opened Dec 16, 2020

• [Build] [ROCm] [Bazel] /opt/rocm/.info/version-dev file not found

  #45733 opened Dec 16, 2020

• ambiguity risk of crop_size

  #45731 opened Dec 16, 2020

• How to specify `force_gpu_compatible` in TensorFlow 2.x?

  #45730 opened Dec 16, 2020

• Issues about the TRTEngineOp

  #45727 opened Dec 16, 2020

• Different fill value for different channel with ImageProjectiveTransformV3

  #45717 opened Dec 15, 2020

• Model training stalls forever after just a few batches.

  #45716 opened Dec 15, 2020

• change LSTM to CuDNNLSTM cause core dumped

  #45709 opened Dec 15, 2020

• Validation Split causing ResourceExhaustedError: OOM when allocating tensor with shape[54000,128,128,3]

  #45708 opened Dec 15, 2020

• iOS Build TFlite as framework

  #45707 opened Dec 15, 2020

• Dot layer incomplete description

  #45706 opened Dec 15, 2020

• Wrong default value in GRU layer documentaion

  #45705 opened Dec 15, 2020

• Wrong error message for DepthwiseConv2D

  #45703 opened Dec 15, 2020

• Why TPU uses more memory than GPU?

  #45702 opened Dec 15, 2020

• To provide the prediction probabilities for Binary classification task using the sigmoid activation in the output layer.

  #45701 opened Dec 15, 2020

• Request for a tutorial on building C++ API not pure C

  #45700 opened Dec 15, 2020

• Failed to run Benchmark

  #45698 opened Dec 15, 2020

• micro: Port TRANSPOSE from lite to micro

  #45695 opened Dec 15, 2020

• micro: port ops BATCH_TO_SPACE_ND and SPACE_TO_BATCH_ND from lite

  #45693 opened Dec 15, 2020

• Unable to compile the TFLite for Microcontrollers C++ Library on Raspberry Pi

  #45692 opened Dec 15, 2020

• Problem with installing Tensorflow on Anaconda

  #45688 opened Dec 15, 2020

• StringLookup on GPU and disable_eager_mode gives incosistent results

  #45687 opened Dec 15, 2020

• ERROR: flatbuffers::flatbuffer_version_string Multiply defined Global Symbol

  #45686 opened Dec 15, 2020

• tensorflow_text.BertTokenizer not giving correct offset values in strings that have special characters

  #45684 opened Dec 15, 2020

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

  #45681 opened Dec 15, 2020

• ERROR: C:/pathtensor/tensorflow-avx2/tensorflow/python/BUILD:5771:1: C++ compilation of rule '//tensorflow/python:_tf_stack.so' failed (Exit 2): python.exe failed: error executing command

  #45680 opened Dec 15, 2020

• Tensorflow 2.4 takes 3 seconds per epoch during training versus 1 second with TensorFlow 2.3

  #45676 opened Dec 14, 2020

• Default Installation is causing errors

  #45671 opened Dec 14, 2020

• Unable to read pb file in OpenCV or Tensorflow in Python

  #45670 opened Dec 14, 2020

• Multi-GPU selection in map_fn and vectorized_map

  #45668 opened Dec 14, 2020

• Out of memory in some tests due to GPU memory limit confusion

  #45664 opened Dec 14, 2020

• No registered 'ResourceScatterNdUpdate' OpKernel for 'GPU'

  #45663 opened Dec 14, 2020

• Wrong device returned for GPUCompatibleFIFOQueueTests.testEnqueueDequeue test

  #45662 opened Dec 14, 2020

• Test failures with "OpError not raised"

  #45661 opened Dec 14, 2020

• Seg faults in various tests

  #45660 opened Dec 14, 2020

• Invalid argument: During Variant Host->Device Copy: non-DMA-copy attempted of tensor type: string

  #45659 opened Dec 14, 2020

• Error during training: F tensorflow/core/common_runtime/gpu/gpu_event_mgr.cc:273] Unexpected Event status: 1

  #45658 opened Dec 14, 2020

• micro: port op FLOOR_DIV from lite

  #45657 opened Dec 14, 2020

• sum/avg/mean ops of tf.keras.backend do not work correctly over batch axis

  #45656 opened Dec 14, 2020

• Mask RCNN tflite inference on android

  #45651 opened Dec 14, 2020

• W/native: op_kernel.cc:1401 OP_REQUIRES failed at save_restore_tenProplem restoring the checkpoint in my android app: sor.cc:175 : Not found: Unsuccessful TensorSliceReader constructor: Failed to find any matching files for [B@66ce18

  #45650 opened Dec 14, 2020

• Support `class_weight` for TimeSeries models in Keras (`model.fit`)

  #45649 opened Dec 14, 2020

• 2.4.0 sess.run freezing in Windows 10 with RTX 3090

  #45648 opened Dec 14, 2020

• Can't Install TF on Apple M1

  #45645 opened Dec 14, 2020

• Tensorflow 1.x binary not found with pip install (Python=3.7)

  #45642 opened Dec 13, 2020

• Is input_shape on Dense for public use?

  #45637 opened Dec 13, 2020

• tf.enable_eager_execution() not working

  #45636 opened Dec 13, 2020

• RTX 3060 TI, creating GPU device take about 5 Minutes

  #45635 opened Dec 13, 2020

• Mapping LSTM over arbitrary tensor shapes

  #45634 opened Dec 13, 2020

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

  #45633 opened Dec 13, 2020

• Conda install Tensorflow 2.1 with mkl (avx/avx2 support) not working

  #45632 opened Dec 12, 2020

• Help with object detection project

  #45630 opened Dec 12, 2020

• The current TensorFlow lite version cannot be built with Visual Studio 2017.

  #45629 opened Dec 12, 2020

• Potential GPU Memory Leak

  #45627 opened Dec 12, 2020

• WARNING:tensorflow:AutoGraph could not transform <function Model.make_predict_function.<locals>.predict_function at 0x0000028717BD2C18> and will run it as-is. Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause:

  #45626 opened Dec 12, 2020

• How to set the signature of keras.models.save_model

  #45625 opened Dec 12, 2020

• tf.test.gpu_device_name() breaks tensorflow

  #45622 opened Dec 12, 2020

• tensorflow.python.framework.errors_impl.UnknownError: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.

  #45616 opened Dec 11, 2020

• metrics in the Classification on imbalanced data tutorial

  #45615 opened Dec 11, 2020

• Tensorflow crashes python when using Keras Convolution/Max Pooling layers on rtx 3070

  #45611 opened Dec 11, 2020

• Tensorflow_datsaet api fails to cache my training data.

  #45609 opened Dec 11, 2020

• Micro: port op CAST from Lite

  #45608 opened Dec 11, 2020

• TFLM: Adding ported and optimized operations for CEVA-BX1

  #45607 opened Dec 11, 2020

• No documentation on how to access the ram:// filesystem of TPU's

  #45605 opened Dec 11, 2020

• No gradients available error

  #45604 opened Dec 11, 2020

• Cannot convert numpy array to tensor in Image Data Generator

  #45601 opened Dec 11, 2020

• Please let tf.keras.metrics.MeanIoU support y_pred from logits

  #45598 opened Dec 11, 2020

• CancelledError: [_Derived_]RecvAsync is cancelled.

  #45594 opened Dec 11, 2020

• Keras doesn't release the memory after fit (single thread and multiprocessing)

  #45592 opened Dec 11, 2020

• Value 'sm_86' is not defined for option 'gpu-name'

  #45590 opened Dec 11, 2020

• Inconsistent results when using tf_upgrade_v2 script

  #45585 opened Dec 10, 2020

• Autograph warning seems to lead to reduced performance

  #45584 opened Dec 10, 2020

• "SAME" padding for avg_pool2d yields different results than explicitly 0-padded input with "VALID"

  #45579 opened Dec 10, 2020

• Tutorials still mention boosted_trees_classifier_train_in_memory

  #45577 opened Dec 10, 2020

• Error when building tensorflow from source

  #45575 opened Dec 10, 2020

• TF 2.4rc3 fails with CUDNN_STATUS_INTERNAL_ERROR on GeForce RTX 3070

  #45574 opened Dec 10, 2020

• save_model/load_model pair does not preserve model's metric

  #45573 opened Dec 10, 2020

• A quantized tf code that contain tf.where op cannot be converted to tflite (operator of type Where is not yet implemented.)

  #45572 opened Dec 10, 2020

• NCCL version is always 2.7.3

  #45571 opened Dec 10, 2020

• failed to compile tensorflow (V2.4_RC4) in debug mode

  #45570 opened Dec 10, 2020

• Add RaggedTensor support to tf.stop_gradient and tf.reduce_logsumexp

  #45569 opened Dec 10, 2020

• Using mixed precision causes incorrect loss, metric values

  #45568 opened Dec 10, 2020

• raise ValueError("Shapes %s and %s are incompatible" % (self, other)) ValueError: Shapes (2, 56, 56, 1024) and () are incompatible

  #45566 opened Dec 10, 2020

• Keras optimizer uses non-trainable variables instead of constants for hyperparameters

  #45558 opened Dec 9, 2020

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

  #45557 opened Dec 9, 2020

• lstm related tests are failing due to unprovided optional inputs are not read correctly as nullptr on s390x machine

  #45556 opened Dec 9, 2020

• Tensorflow does not use the GPU during training despite manual activation

  #45546 opened Dec 9, 2020

• Windows Tensorflow load large than 4GB file will meet outOfRange error.

  #45543 opened Dec 9, 2020

• Official Example For TensorFlow Lite Model Maker keeps crashing

  #45541 opened Dec 9, 2020

• Bug in tensorflow.python.ops.parallel_for.gradients.batch_jacobian()

  #45539 opened Dec 9, 2020

277 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.

• Add op to convert numeric tensor to string faster

  #45466 commented on Dec 10, 2020 • 11 new comments

• Use copy instead of deepcopy for sklearn wrapper

  #41341 commented on Dec 10, 2020 • 9 new comments

• Add GPU implementation for tf.segment_mean

  #44911 commented on Dec 14, 2020 • 9 new comments

• TF Lite issue when loading a saved TF Lite model on platforms with different endianness

  #45009 commented on Dec 13, 2020 • 7 new comments

• maskrcnn convert error

  #45199 commented on Dec 14, 2020 • 7 new comments

• While running tenosrflow object detection code i got this error.

  #45422 commented on Dec 16, 2020 • 7 new comments

• Refactor exp.h from reference_ops.h; 2nd step to port op EXP for TFLi…

  #45535 commented on Dec 15, 2020 • 7 new comments

• Layer.add_loss has wrong behaviour when building Sequential model in for loop

  #44590 commented on Dec 11, 2020 • 6 new comments

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

  #45385 commented on Dec 12, 2020 • 5 new comments

• How to implement data-dependent initialization with MultiGPU/TPU [Tensorflow >= 2.x]

  #39645 commented on Dec 13, 2020 • 5 new comments

• Segmenation fault when saving StringLookup layer

  #45449 commented on Dec 15, 2020 • 5 new comments

• save/load pair do not preserve `trainable` attribute

  #45537 commented on Dec 15, 2020 • 5 new comments

• tflite 2.4 build benchmark_model error on ubuntu 16.04 (cmake)

  #45290 commented on Dec 16, 2020 • 5 new comments

• TF-TRT slice op converter dynamic shape

  #40736 commented on Dec 12, 2020 • 5 new comments

• Do not call tf.shape on a ragged tensor when computing losses.

  #45015 commented on Dec 16, 2020 • 5 new comments

• Make TensorFlow build on Apple Silicon

  #45404 commented on Dec 10, 2020 • 5 new comments

• TF2 GPUs on Kubernetes

  #45494 commented on Dec 9, 2020 • 4 new comments

• Enable use of generators in tf lite model maker

  #45378 commented on Dec 12, 2020 • 4 new comments

• If TMP is not set, then the build system defaults to the wrong path

  #40614 commented on Dec 14, 2020 • 4 new comments

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

  #45313 commented on Dec 16, 2020 • 4 new comments

• Bus error when load the tflite model

  #45504 commented on Dec 16, 2020 • 4 new comments

• enlarge the format data width

  #43587 commented on Dec 15, 2020 • 4 new comments

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

  #43900 commented on Dec 14, 2020 • 4 new comments

• Avoid doing reset when position is still in buffer in BufferedInputStream.

  #44104 commented on Dec 16, 2020 • 4 new comments

• update grpc to v1.32.0 to fix val.val issues while compiling TFS from source

  #44282 commented on Dec 14, 2020 • 4 new comments

• [XLA] Lower TF/XLA overhead by creating less objects

  #45022 commented on Dec 11, 2020 • 4 new comments

• TFLu: Port squeeze op from TFLite

  #45144 commented on Dec 16, 2020 • 4 new comments

• [XLA] Fix regression for small element-wise kernels.

  #45343 commented on Dec 11, 2020 • 4 new comments

• Suspected memory leak - when loading multiple models with tf.saved_model.load

  #45453 commented on Dec 10, 2020 • 3 new comments

• Fail to run tflite on DSP - BATCH_TO_SPACE_ND failed to prepare

  #45478 commented on Dec 10, 2020 • 3 new comments

• docs: Could not load dynamic library 'libcublas.so.10'

  #44312 commented on Dec 11, 2020 • 3 new comments

• MUL doesn't work on gpu delegate with error: Dimension can not be reduced to linear

  #37102 commented on Dec 14, 2020 • 3 new comments

• Custom loop training of tacotron2 with MultiWorkerMirroredStrategy - Exceptions and failures when use MultiWorkerMirroredStrategy -

  #45056 commented on Dec 14, 2020 • 3 new comments

• WARNING:tensorflow:AutoGraph could not transform <function canonicalize_signatures.<locals>.signature_wrapper at 0x7ff4c27f98c0> and will run it as-is.

  #45354 commented on Dec 14, 2020 • 3 new comments

• Incorrect instructions for Tensor RT

  #45410 commented on Dec 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 Dec 15, 2020 • 3 new comments

• Functionality of writing back Tflite model with multiple subgraphs that is currently loaded in the Tflite Interpreter.

  #44606 commented on Dec 15, 2020 • 3 new comments

• [TFLite] TF MatMul operator converted to an unsupported TFL 16x16 FullyConnected operator with 16x8 post-training quantization

  #44707 commented on Dec 15, 2020 • 3 new comments

• Optimizing a custom loss in TF1 vs TF2

  #45186 commented on Dec 15, 2020 • 3 new comments

• TFLite Converter for BiLSTM based architecture

  #45361 commented on Dec 15, 2020 • 3 new comments

• tf.while_loop how to create a nested loop

  #45337 commented on Dec 15, 2020 • 3 new comments

• cp: cannot stat '/usr/include/x86_64-linux-gnu/NvInferPlugin.h': No such file or directory

  #37341 commented on Dec 15, 2020 • 3 new comments

• Selective builds to reduce tensorflow-lite binary size missing needed java package org.tensorflow.lite

  #45488 commented on Dec 15, 2020 • 3 new comments

• GeForce 3090 incompatibility with Nightly

  #44959 commented on Dec 16, 2020 • 3 new comments

• SubProcess ended with return code: 4294967295 with tensorflow-gpu 2.4.0rc cuda11.0 cudnn8.0.2 on windows 10

  #45170 commented on Dec 16, 2020 • 3 new comments

• Fix slim record_writer compression type

  #44786 commented on Dec 14, 2020 • 3 new comments

• [INTEL MKL] Added mkl_einsum op

  #45118 commented on Dec 16, 2020 • 3 new comments

• [pylintrc] Enforce Google docstring format

  #45427 commented on Dec 16, 2020 • 3 new comments

• Refactor Transpose reference op for port to TFLu

  #45438 commented on Dec 15, 2020 • 3 new comments

• Add gradient for QR decomposition of complex matrices.

  #45475 commented on Dec 11, 2020 • 3 new comments

• Model converted to TFLite always returns NaN as output.

  #22803 commented on Dec 9, 2020 • 2 new comments

• TFLite model output different probabilities and thus different classes in Android compared to the original h5 model..

  #35538 commented on Dec 10, 2020 • 2 new comments

• Model saving using the tf format uses the wrong TensorSpec for convolutions

  #45077 commented on Dec 11, 2020 • 2 new comments

• Cannot convert model: Exception has occurred: ValueError

  #45460 commented on Dec 11, 2020 • 2 new comments

• tensorflow.python.framework.errors_impl.InvalidArgumentError: 'func' argument to TF_GraphCopyFunction cannot be null

  #44403 commented on Dec 11, 2020 • 2 new comments

• Tensorboard hparams plugin API reports final epoch metrics, val loss, *not* from epoch with min val loss

  #45384 commented on Dec 11, 2020 • 2 new comments

• Met unsupported operator of type Cast when quantize model get from quantize aware training

  #26478 commented on Dec 12, 2020 • 2 new comments

• Tensor shape signature is not parsed safely with `ParseTensors` of TF Lite InterpreterBuilder on Big Endian machines

  #45210 commented on Dec 12, 2020 • 2 new comments

• tf.while_loop allow to vary shapes by default

  #45538 commented on Dec 12, 2020 • 2 new comments

• Groups parameter of Conv2d and Conv2dTranpose ('deconvolution') not Working ?

  #45216 commented on Dec 12, 2020 • 2 new comments

• Dll load failed error

  #45398 commented on Dec 13, 2020 • 2 new comments

• Windows - tensorflow.python.framework.errors_impl.UnknownError: Failed to rename:

  #41380 commented on Dec 13, 2020 • 2 new comments

• DataType error: DataType 6 is not recognized in Java (version 2.4.0)

  #41212 commented on Dec 14, 2020 • 2 new comments

• Decoding output of object detection mobile_ssd_v2_float_coco.tflite

  #40298 commented on Dec 14, 2020 • 2 new comments

• [TF 2.4] KerasTensor breaks typing compatibility

  #44613 commented on Dec 14, 2020 • 2 new comments

• Confusion regarding TensorFlow licensing

  #41739 commented on Dec 14, 2020 • 2 new comments

• mixed_float16 + example code = horrible performance on a GeForce GTX 1660 Ti

  #44191 commented on Dec 14, 2020 • 2 new comments

• Same code - difference result from PC and jetson nano

  #45483 commented on Dec 15, 2020 • 2 new comments

• error in converting ssdlite_mobilenet_v2_coco_2018_05_09 tflite_graph.pb to tflite

  #45302 commented on Dec 15, 2020 • 2 new comments

• Converted TensorFlow Lite model is not a valid model in Android Studio

  #45373 commented on Dec 15, 2020 • 2 new comments

• Tensorflow/Keras Endpoint layer does not work with Tensorflow dataset in eager mode

  #45433 commented on Dec 15, 2020 • 2 new comments

• Source compilation fails on Musl system on multiple places

  #45446 commented on Dec 15, 2020 • 2 new comments

• RTX 3070 - CUBLAS_STATUS_ALLOC_FAILED

  #45383 commented on Dec 15, 2020 • 2 new comments

• Cannot save a subclassed keras model that relies on AutoGraph

  #44430 commented on Dec 15, 2020 • 2 new comments

• Tensorflow-Python Environment Assertion Error

  #45396 commented on Dec 16, 2020 • 2 new comments

• Use of GPU delegate increases inference runtime in Android's Native environment (C++) but not using Java API

  #42709 commented on Dec 16, 2020 • 2 new comments

• CategoryEncoding not working when loading model again

  #41552 commented on Dec 16, 2020 • 2 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 Dec 16, 2020 • 2 new comments

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

  #39452 commented on Dec 14, 2020 • 2 new comments

• Enforce host memory for DT_INT32 inputs of functional ops

  #39860 commented on Dec 16, 2020 • 2 new comments

• TypeError on line 88 should have a condition

  #43260 commented on Dec 16, 2020 • 2 new comments

• Update math_ops.py

  #44902 commented on Dec 16, 2020 • 2 new comments

• [ROCm] Adding new CI scripts for ROCm platform

  #45159 commented on Dec 15, 2020 • 2 new comments

• [ROCm] Adding ROCm support for the `run_pip_tests.sh` script

  #45161 commented on Dec 15, 2020 • 2 new comments

• [ROCm] Adding / removing no_rocm tag from some unit tests

  #45166 commented on Dec 15, 2020 • 2 new comments

• [XLA] Fix a dangling reference bug that is triggered in Ubuntu 20.04 with GCC 9.3

  #45275 commented on Dec 16, 2020 • 2 new comments

• [ROCm] Populating memory bandwidth information in the DeviceDescription

  #45416 commented on Dec 15, 2020 • 2 new comments

• [ROCm] Removing --action_env=ROCM_ROOT... when building with --config=rocm

  #45456 commented on Dec 15, 2020 • 2 new comments

• report error "tensorflow.python.framework.errors_impl.InvalidArgumentError "when visit dataset

  #45532 commented on Dec 9, 2020 • 1 new comment

• ESP32 LyraT: Audio data/I2S issue

  #45185 commented on Dec 9, 2020 • 1 new comment

• TFLite GPU execution failed

  #35632 commented on Dec 9, 2020 • 1 new comment

• Calling tf.function from tf.py_function in dataset.map hangs.

  #32454 commented on Dec 9, 2020 • 1 new comment

• Tensorflow lite C++ standalone projects from 'generate_projects' missing sparkfun headers?

  #35096 commented on Dec 9, 2020 • 1 new comment

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

  #43800 commented on Dec 9, 2020 • 1 new comment

• TensorFlow Lite for Microcontrollers fails to build

  #34664 commented on Dec 9, 2020 • 1 new comment

• [TF 2.0] allow tf.function input_signature to be specified by annotations

  #31579 commented on Dec 9, 2020 • 1 new comment

• AUTHORS references non-existent file

  #28036 commented on Dec 9, 2020 • 1 new comment

• ValueError when computing Jacobian in eager mode

  #34260 commented on Dec 9, 2020 • 1 new comment

• [API] Tensorflow.Tensor.Tensor(byte[ ][ ], long[ ]) constructor not exist any more

  #44225 commented on Dec 9, 2020 • 1 new comment

• Different ABI between docker version and public version in tf 2.3.0

  #44255 commented on Dec 9, 2020 • 1 new comment

• Much worse performance when using mixed precision training (using tensorflow.keras policy)

  #39556 commented on Dec 10, 2020 • 1 new comment

• RuntimeError: Regular TensorFlow ops are not supported by this interpreter. Make sure you apply/link the Flex delegate before inference.Node number 51 (FlexResizeNearestNeighbor) failed to prepare.

  #45143 commented on Dec 10, 2020 • 1 new comment

• Segmentation fault during training simple decision tree classifier

  #45310 commented on Dec 10, 2020 • 1 new comment

• distributed_dataset tensorflow.python.framework.errors_impl.UnavailableError: Socket closed

  #43362 commented on Dec 10, 2020 • 1 new comment

• No rule to make target 'hello_world_bin'. Stop

  #44841 commented on Dec 10, 2020 • 1 new comment

• how to understand handling val data with keras in tf2.1

  #45357 commented on Dec 10, 2020 • 1 new comment

• Keras EarlyStopping callback on val_auc running mysteriously

  #43645 commented on Dec 10, 2020 • 1 new comment

• Update Keras Conv2D layer docs to specify padding value

  #43886 commented on Dec 10, 2020 • 1 new comment

• Errors when loading a model having keras.Lambda

  #45180 commented on Dec 10, 2020 • 1 new comment

• InvalidArgumentError: assertion failed: [predictions must be <= 1] [Condition x <= y did not hold element-wise:] [x (model/dense_3/Relu:0) = ] [[0.179184318 1.62250423 0...]...] [y (metrics/auc/Cast_2/x:0) = ] [1] [[{{node metrics/auc/assert_less_equal/Assert/AssertGuard/else/_11/Assert}}]] [Op:__inference_distributed_function_8508]

  #45241 commented on Dec 10, 2020 • 1 new comment

• Missing headers when using the C++ library

  #42492 commented on Dec 10, 2020 • 1 new comment

• Install tensorflow in python 3

  #44546 commented on Dec 10, 2020 • 1 new comment

• Incompatibility between versions of TF and CUDA dynamic libraries

  #44159 commented on Dec 10, 2020 • 1 new comment

• Unbalanced workload distribution on parameter server replicas in distributed tensorflow (parameter-server strategy)

  #45291 commented on Dec 10, 2020 • 1 new comment

• SeparableConv documention missing argument constraint

  #45259 commented on Dec 10, 2020 • 1 new comment

• training keras Sequential model with python generator raise error: AttributeError: 'tuple' object has no attribute 'rank'

  #44285 commented on Dec 11, 2020 • 1 new comment

• On-disk cache for compiled functions

  #44710 commented on Dec 11, 2020 • 1 new comment

• [mbed] ERROR: Unable to write build ignore file

  #45292 commented on Dec 11, 2020 • 1 new comment

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

  #44021 commented on Dec 11, 2020 • 1 new comment

• Missing call to model.optimizer._clip_gradients(grads)?

  #45146 commented on Dec 11, 2020 • 1 new comment

• If-Condition with TF-Bools looses track of namescopes

  #36464 commented on Dec 11, 2020 • 1 new comment

• [FIX ME!] Matrix computaion in custom-ops got partly correct result.

  #44945 commented on Dec 11, 2020 • 1 new comment

• load_model in tf>=2.3.0 raises TypeError when loading a model saved in tf 2.2.0

  #45288 commented on Dec 11, 2020 • 1 new comment

• Could not compute output tensor even when correct inputs are being passed to keras model

  #44766 commented on Dec 11, 2020 • 1 new comment

• randomly discontinuation when model training with tf.keras in tf-gpu-2.1

  #45297 commented on Dec 11, 2020 • 1 new comment

• docs: Failed to initialize NVML: Driver/library version mismatch

  #44301 commented on Dec 11, 2020 • 1 new comment

• Segfault with tf.linalg.cholesky

  #43170 commented on Dec 11, 2020 • 1 new comment

• KeyError on concrete_functions while loading a model

  #42004 commented on Dec 11, 2020 • 1 new comment

• I had tried this code in Xcode with objective c but I am getting the error "Explicit specialization of undeclared template struct 'NumTraits'

  #44288 commented on Dec 11, 2020 • 1 new comment

• BrokenPipeError: [Errno 32] Broken pipe

  #44453 commented on Dec 11, 2020 • 1 new comment

• RuntimeError: The layer has never been called and thus has no defined output shape.

  #45363 commented on Dec 11, 2020 • 1 new comment

• Support Python 3.9

  #44485 commented on Dec 11, 2020 • 1 new comment

• Inconsistency in loss on SAME data for train and validation modes

  #44568 commented on Dec 11, 2020 • 1 new comment

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

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

• Add Ubuntu 20.04 installation instructions

  #43693 commented on Dec 11, 2020 • 1 new comment

• could I build both amd and nvidia support?

  #45367 commented on Dec 11, 2020 • 1 new comment

• how to handle non-tail recursive in tf when using control flow

  #45165 commented on Dec 11, 2020 • 1 new comment

• "NotImplementedError: Missing functionality to restore state of object" when using StringLookup() and CategoryEncoding()

  #44872 commented on Dec 11, 2020 • 1 new comment

• how optimize a noise image with tensorflow?

  #45284 commented on Dec 11, 2020 • 1 new comment

• tf.keras SparseCategoricalCrossentropy with sample_weight on TPU: Error DenseToDenseSetOperation (No registered 'DenseToDenseSetOperation)

  #43195 commented on Dec 11, 2020 • 1 new comment

• Docs for install cuda 11.0 and cudnn 8

  #45308 commented on Dec 11, 2020 • 1 new comment

• TensorFlow 1.15.4 artifacts on Maven Central?

  #45370 commented on Dec 11, 2020 • 1 new comment

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

  #43527 commented on Dec 11, 2020 • 1 new comment

• tensorflow.keras MirroredStrategy hangs without error message

  #45178 commented on Dec 12, 2020 • 1 new comment

• yolov3-tiny tflite model of shape [1, 2535, 85] representation

  #44890 commented on Dec 12, 2020 • 1 new comment

• AttributeError: 'TensorArray' object has no attribute 'mark_used' with tf.function

  #45093 commented on Dec 12, 2020 • 1 new comment

• Downloading pretrained Efficient Det in google colab using TensorFlow Object Detection Api gives a series of unknown warnings?

  #44161 commented on Dec 12, 2020 • 1 new comment

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

  #42506 commented on Dec 12, 2020 • 1 new comment

• tf.nn.max_pool3d crashes(floating point exception) when strides=0

  #42213 commented on Dec 12, 2020 • 1 new comment

• When tf.image.central_crop taking a tensor as `central_fraction` argument, it raises OperatorNotAllowedInGraphError.

  #45324 commented on Dec 13, 2020 • 1 new comment

• [RNN] dynamic input shape support

  #40185 commented on Dec 13, 2020 • 1 new comment

• Failed to convert .pb file to rflite

  #39315 commented on Dec 13, 2020 • 1 new comment

• libmkl_intel.so and libiomp5.so are not being copied for libtensorflow.so build

  #37337 commented on Dec 13, 2020 • 1 new comment

• [CPU] Best CPU Build found with MKL config in bazel build BUT with MKL Disabled in script, especially on LSTM, can't undestand why

  #38858 commented on Dec 13, 2020 • 1 new comment

• Allow --cpu=x64_arm64_windows as a compilation target on Windows

  #45151 commented on Dec 13, 2020 • 1 new comment

• Keras optional Input with default value

  #40997 commented on Dec 14, 2020 • 1 new comment

• Documentation for ensuring CUPTI for Profiling is Misleading

  #39526 commented on Dec 14, 2020 • 1 new comment

• tf.device scope not working correctly

  #44510 commented on Dec 14, 2020 • 1 new comment

• Import Error DLL failed

  #45287 commented on Dec 14, 2020 • 1 new comment

• ImportError: Traceback (most recent call last): File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 64, in <module> from tensorflow.python._pywrap_tensorflow_internal import * ImportError: DLL load failed: The specified module could not be found.

  #45435 commented on Dec 14, 2020 • 1 new comment

• tflite RuntimeError: Encountered unresolved custom op: StridedSlice failed to prepare.

  #45377 commented on Dec 14, 2020 • 1 new comment

• I can build tensorflow.dll successfully,but not include SSE4.2,Why?

  #45421 commented on Dec 14, 2020 • 1 new comment

• cuDNN LSTMs not utilized after loading a model using `tf.saved_model.load`

  #44624 commented on Dec 14, 2020 • 1 new comment

• Logging Memory Usage For each TensorFlow Operation

  #42584 commented on Dec 14, 2020 • 1 new comment

• Tensorflow 2.3.1 is not compatible with cuda 10.2 on Windows

  #44622 commented on Dec 14, 2020 • 1 new comment

• keras.model.predict doesn't work when using multiprocessing

  #45426 commented on Dec 14, 2020 • 1 new comment

• all CUDA-capable devices are busy or unavailable

  #45450 commented on Dec 14, 2020 • 1 new comment

• .ckpt to .pb

  #45341 commented on Dec 14, 2020 • 1 new comment

• ValueError: Input 0 of layer dense_2 is incompatible with the layer: expected axis -1 of input shape to have value 50176 but received input with shape [None, 57600]

  #45362 commented on Dec 14, 2020 • 1 new comment

• libcusolver.so.10?

  #45263 commented on Dec 14, 2020 • 1 new comment

• Use tf.Tensor as input to tf.io.gfile.GFile

  #32620 commented on Dec 14, 2020 • 1 new comment

• [TFLite] Experimental new quantizer calculates the wrong scaling when the input range doesn't include 0

  #45038 commented on Dec 14, 2020 • 1 new comment

• Memory leak when repeatedly loading and deleting keras models

  #40171 commented on Dec 14, 2020 • 1 new comment

• module 'gast' has no attribute 'Ellipsis'

  #45295 commented on Dec 14, 2020 • 1 new comment

• 404 Error during oxford-pets tutorial dataset loading

  #31171 commented on Dec 14, 2020 • 1 new comment

• tensorflow-nightly-gpu looking for cusolver64_10.dll on a cuDNN 11.1 installation

  #44291 commented on Dec 15, 2020 • 1 new comment

• TF 2.3 S3 client having permission issues with `tf.data.TFRecordDataset()`

  #44564 commented on Dec 15, 2020 • 1 new comment

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

  #43911 commented on Dec 15, 2020 • 1 new comment

• Dealing with `from tensorflow.contrib import ...` in tf_upgrade_v2

  #44103 commented on Dec 15, 2020 • 1 new comment

• convert tesnorflow::ops::cast to tensorflow::tensor

  #42588 commented on Dec 15, 2020 • 1 new comment

• Tflite pod lint error

  #45198 commented on Dec 15, 2020 • 1 new comment

• Failed to load native tensorflow runtime

  #45471 commented on Dec 15, 2020 • 1 new comment

• LookupError when computing nested gradient with UpSampling2D

  #43648 commented on Dec 15, 2020 • 1 new comment

• opcode 'SUB' version '3' in tfllite_runtime on Windows

  #44237 commented on Dec 15, 2020 • 1 new comment

• TF Nightly fails to find most recent cusolver64

  #44381 commented on Dec 15, 2020 • 1 new comment

• Tensorflow gpu does not work with RTX 3000 series card.

  #45285 commented on Dec 15, 2020 • 1 new comment

• tf.sparse.reorder crashes Tensorflow if number of possible entries of dense_shape overflows 64 bit int

  #45392 commented on Dec 16, 2020 • 1 new comment

• When pre-built libtensorflow C API will be updated to 2.4?

  #45477 commented on Dec 16, 2020 • 1 new comment

• model.export(export_dir='.') does not export labels nor model.

  #41545 commented on Dec 16, 2020 • 1 new comment

• tf.function-decorated function tried to create variables on non-first call

  #27120 commented on Dec 16, 2020 • 1 new comment

• segfault in FlexDelegate on android

  #38025 commented on Dec 16, 2020 • 1 new comment

• Numerical instability of gradient calculation of tf.norm (nan at 0, inf for small values)

  #12071 commented on Dec 16, 2020 • 1 new comment

• Running a part of a calculation in double precision

  #45482 commented on Dec 16, 2020 • 1 new comment

• Freeze improvements

  #39286 commented on Dec 14, 2020 • 1 new comment

• Add blank_index parameter for ctc_greedy_decoder

  #39324 commented on Dec 14, 2020 • 1 new comment

• [TF-TRT] Fix TopK Combined NMS

  #40062 commented on Dec 16, 2020 • 1 new comment

• TF-TRT Improve reshape op converter

  #40545 commented on Dec 12, 2020 • 1 new comment

• Patch to core for RFC: Sparse Domain Isolation for Supporting Large-scale Sparse Weights

  #41371 commented on Dec 14, 2020 • 1 new comment

• [TFLite 16x8] Fixes for TANH and LOGISTIC

  #42671 commented on Dec 15, 2020 • 1 new comment

• C++ convert_numpy_inputs, _as_ndarray, _is_ndarray

  #42695 commented on Dec 16, 2020 • 1 new comment

• Overview sanilty check fails on pylint upgrade

  #43040 commented on Dec 16, 2020 • 1 new comment

• C++ canonicalize_function_inputs WIP

  #43053 commented on Dec 16, 2020 • 1 new comment

• Fixed inputs not being cast to tuple

  #43774 commented on Dec 14, 2020 • 1 new comment

• [TF-numpy] `imag` and `real` returns scalar when `val` is a scalar.

  #44382 commented on Dec 11, 2020 • 1 new comment

• optimize all_reduce_indexed_slices for better performance

  #44492 commented on Dec 10, 2020 • 1 new comment

• [Dockerfile] move stubs to the end of LD_LIBRARY_PATH

  #44732 commented on Dec 14, 2020 • 1 new comment

• [CostModel] Generating CostGraphDef with Global Index

  #44794 commented on Dec 9, 2020 • 1 new comment

• xrt ops use bfc_allocator instead

  #44985 commented on Dec 14, 2020 • 1 new comment

• Use batch_size without modulo

  #45047 commented on Dec 10, 2020 • 1 new comment

• Fix Const op tensor_content on s390x during save/load

  #45339 commented on Dec 15, 2020 • 1 new comment

• [*] Rename "Arguments:" to "Args:"

  #45420 commented on Dec 16, 2020 • 1 new comment

• "keras predict" uses one more batch than specified by "steps" argument with generator x

  #45459 commented on Dec 9, 2020 • 0 new comments

• Feed batch of images in Java

  #45347 commented on Dec 10, 2020 • 0 new comments

• Tensorflow 2.3.0 TfLite converter adds ExpandDims

  #43141 commented on Dec 10, 2020 • 0 new comments

• STM32F407G Build Error - AddBuiltin() mismatch in arguments

  #42756 commented on Dec 10, 2020 • 0 new comments

• tf.keras.model.Model-subclassed models do not serialise tf.function-ed methods

  #45432 commented on Dec 10, 2020 • 0 new comments

• [RNN] tensorflow/lite/kernels/fully_connected.cc:119 is_optional_bias_int != true error in quantized LSTM model

  #45402 commented on Dec 10, 2020 • 0 new comments

• **kwargs weights - Initialization Layer

  #44767 commented on Dec 10, 2020 • 0 new comments

• Support for CUDA 11.1 and RTX 3090

  #45319 commented on Dec 10, 2020 • 0 new comments

• Memory leak using nested name_scope in eager mode

  #45322 commented on Dec 10, 2020 • 0 new comments

• nnapi delegate and cpu output with my QAT int8 model got Inconsistent result

  #45394 commented on Dec 10, 2020 • 0 new comments

• tf.keras.preprocessing.image.random_shear raises TypeError when it is mapped to a tf dataset.

  #45326 commented on Dec 10, 2020 • 0 new comments

• Compilation Problems

  #45405 commented on Dec 10, 2020 • 0 new comments

• Different variable names in keras layers depending on how layer is used

  #29576 commented on Dec 10, 2020 • 0 new comments

• Incorrect Layer Normalization description

  #39942 commented on Dec 10, 2020 • 0 new comments

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

  #40501 commented on Dec 10, 2020 • 0 new comments

• Insert Windows command line in the tutorials

  #33923 commented on Dec 10, 2020 • 0 new comments

• Exception: <unknown>:0: error: loc("batch_normalization/moving_mean"): is not immutable, try running tf-saved-model-optimize-global-tensors to prove tensors are immutable

  #44790 commented on Dec 10, 2020 • 0 new comments

• postnet android

  #44821 commented on Dec 10, 2020 • 0 new comments

• TimeDistributed Layer Applied to Masked Layer (Functional API) not working under TF 2.3.1

  #45371 commented on Dec 10, 2020 • 0 new comments

• Op to convert numeric tensor to string type faster

  #45406 commented on Dec 10, 2020 • 0 new comments

• TF-TRT Dynamic Shapes Feature Tracker

  #45481 commented on Dec 10, 2020 • 0 new comments

• "error: 'tf.Conv2D' op is neither a custom op nor a flex op" when converting a basic MNIST network

  #45252 commented on Dec 11, 2020 • 0 new comments

• Training takes forever when applying distribute strategy

  #44698 commented on Dec 11, 2020 • 0 new comments

• Problems to convert a tensorflow model to TFLite

  #44918 commented on Dec 11, 2020 • 0 new comments

• ambiguous template instantiation error with Eigen code

  #44626 commented on Dec 11, 2020 • 0 new comments

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

  #43942 commented on Dec 11, 2020 • 0 new comments

• Loading a TF1 Protocol buffer does not work in TF versions 2.2.0 and above

  #42980 commented on Dec 11, 2020 • 0 new comments

• Significant accuracy drop during conversion of DistillBert.

  #43476 commented on Dec 11, 2020 • 0 new comments

• Can't change model attribute inside @tf.function with custom training loop.

  #41760 commented on Dec 11, 2020 • 0 new comments

• tf.random.categorical overflow before OOM on GPU

  #41458 commented on Dec 11, 2020 • 0 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 Dec 11, 2020 • 0 new comments

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

  #44005 commented on Dec 11, 2020 • 0 new comments

• recorded_allocation parameters for quantized tensors are incorrect for BE machines

  #45455 commented on Dec 11, 2020 • 0 new comments

• tflite quantized model unable to run on DSP

  #45452 commented on Dec 11, 2020 • 0 new comments

• Tensorflow profiler crashes when using string categorical layer

  #41244 commented on Dec 11, 2020 • 0 new comments

• mnasnet with squeeze and excite can not be quantized

  #45082 commented on Dec 11, 2020 • 0 new comments

• NotFoundError: No algorithm worked! when using Conv2D

  #43174 commented on Dec 11, 2020 • 0 new comments

• Have similar default values for min_delta in ReduceLROnPlateau and EarlyStopping callbacks

  #43100 commented on Dec 11, 2020 • 0 new comments

• Why tf.concat not support multiply types

  #34365 commented on Dec 11, 2020 • 0 new comments

• [TF 2.2] Elapsed time of ConcreteFunction becomes shorter when printing loss

  #40659 commented on Dec 11, 2020 • 0 new comments

• test with random fail on graph mode

  #44642 commented on Dec 12, 2020 • 0 new comments

• Create makefile targets to expose file lists to enable separate scripting

  #45086 commented on Dec 14, 2020 • 0 new comments

• How to unset changes made by tf.debugging.set_log_device_placement?

  #45465 commented on Dec 15, 2020 • 0 new comments

• individual neuron pruning (similar to tfkerassurgeon)

  #42688 commented on Dec 15, 2020 • 0 new comments

• Make callbacks serializable

  #36635 commented on Dec 14, 2020 • 0 new comments

• Added ResNext models

  #37146 commented on Dec 14, 2020 • 0 new comments

• Completed validation_data entry for .fit()

  #38979 commented on Dec 14, 2020 • 0 new comments

• tf.function ensures type annotations and input_signature match

  #40817 commented on Dec 13, 2020 • 0 new comments

• Type inference using type annotations

  #41256 commented on Dec 13, 2020 • 0 new comments

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

  #42373 commented on Dec 16, 2020 • 0 new comments

• Resnet gradients cpp

  #42668 commented on Dec 13, 2020 • 0 new comments

• XLA Asynchronous compilation

  #43034 commented on Dec 11, 2020 • 0 new comments

• [PluggableDevice] Add device alias support

  #43611 commented on Dec 16, 2020 • 0 new comments

• Add GPU headers to //tensorflow/core:headers

  #44370 commented on Dec 12, 2020 • 0 new comments

• support SparseToDenseOp on GPU

  #44702 commented on Dec 14, 2020 • 0 new comments

• [Intel MKL] Enable MklConv/MklFusedConv with explicit padding

  #44934 commented on Dec 14, 2020 • 0 new comments

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

  #44950 commented on Dec 15, 2020 • 0 new comments

• Keras TB callback: Fix write_steps_per_second when profile_batch=0

  #45204 commented on Dec 10, 2020 • 0 new comments

• [XLA] Relaxing requirements for CloneWithNewOperands to preserve sharding info

  #45276 commented on Dec 14, 2020 • 0 new comments

• Update _inferred_steps when recreating iterator.

  #45338 commented on Dec 11, 2020 • 0 new comments

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

  #45342 commented on Dec 13, 2020 • 0 new comments

• Add int8 and int16 support for FILL operator

  #45366 commented on Dec 9, 2020 • 0 new comments

• Optimize MutableDenseHashTable's memory usage and performance in case of large value_shape

  #45407 commented on Dec 9, 2020 • 0 new comments

• Flatbuffer_conversion changes for Transpose

  #45439 commented on Dec 9, 2020 • 0 new comments

• Extract reference for operator DIV to standalone header

  #45442 commented on Dec 14, 2020 • 0 new comments

• micro: copy operator FILL kernel from lite

  #45457 commented on Dec 9, 2020 • 0 new comments