Weekly GitHub Report for Tensorflow - 2024-07-05 19:46:13
Weekly GitHub Report for Tensorflow
Thank you for subscribing to our weekly newsletter! Each week, we deliver a comprehensive summary of your GitHub project's latest activity right to your inbox, including an overview of your project's issues, pull requests, contributors, and commit activity.
I. Issues
1.1 Open Issues
Open Issues This Week: 27
Summarized Issues:
- Installation Issues on Windows: Users are experiencing problems installing TensorFlow on Windows platforms, leading to ImportErrors and other installation failures. These issues often arise due to DLL load failures or SSL certificate verification problems. Despite following various guides, users are unable to successfully integrate TensorFlow into their environments.
- TensorFlow Lite Build and Runtime Errors: Multiple issues have been reported regarding TensorFlow Lite, including build failures on various platforms and runtime errors. Users face problems such as missing dependencies, undefined references, and segmentation faults when running models. These issues affect both Linux and Android platforms.
- github.com/tensorflow/tensorflow/issues/70659
- github.com/tensorflow/tensorflow/issues/70664
- github.com/tensorflow/tensorflow/issues/70730
- github.com/tensorflow/tensorflow/issues/70802
- github.com/tensorflow/tensorflow/issues/70842
- github.com/tensorflow/tensorflow/issues/70851
- github.com/tensorflow/tensorflow/issues/70868
- Python Interpreter and Version Compatibility: Users are encountering issues with TensorFlow's compatibility with different Python interpreters and versions. Problems include TensorFlow using an unexpected Python interpreter during Bazel tests and ImportErrors when using specific TensorFlow functions or libraries.
- Model and Operation Errors: Several issues have been reported related to TensorFlow operations and model configurations. These include bugs in matrix multiplication, errors in model loading, and issues with specific operations like
tf.matmulandtf.keras.metrics.OneHotIoU.
- Inference and Performance Optimization: Users are seeking guidance on optimizing TensorFlow inference performance and managing device-specific operations. Issues include assigning operations to specific devices and handling multiple inputs and outputs in TensorFlow Lite.
- Data Augmentation and Model Training: There are inquiries and issues related to data augmentation techniques and model training processes. Users are looking for effective ways to augment data for object detection and encountering errors during model fine-tuning.
- Documentation and Advisory Errors: Users have reported errors in TensorFlow's documentation and GitHub advisories. These include incorrect proof-of-concept codes in advisories and outdated manylinux support documentation.
- Platform-Specific Issues: Various platform-specific issues have been reported, including TensorFlow not detecting GPUs, build errors on specific platforms, and inaccurate measurements on iOS devices. These issues highlight the challenges of running TensorFlow on diverse hardware and software environments.
- Miscellaneous Bugs and Errors: Other miscellaneous issues include segmentation faults, attribute errors, and TPU resolution errors. These bugs affect various TensorFlow functionalities and environments, causing disruptions in model training and deployment.
- Spam and Irrelevant Posts: Occasionally, spam posts appear on the TensorFlow GitHub repository, which are irrelevant to the project's issues and discussions. These posts need to be identified and removed to maintain the quality of the issue tracker.
1.2 Top 5 Active Issues:
We consider active issues to be issues that have generated much discussion in the issue's comments.
- Cannot take the length of shape with unknown rank.: This issue is about a user encountering a
ValueErrorwhen trying to train a U-Net model using TensorFlow, specifically due to an error related to taking the length of a shape with an unknown rank. The user has shared their training code and the error traceback, seeking help to resolve the problem.
- A commenter suggests that the
build_unetmethod should accept two arguments: the shape of the input as a tuple and the number of dimensions, and provides a code snippet to illustrate the correct usage.- Number of comments: 17
-
Data Service Dynamic sharding policy seems extremely slow: This issue describes a performance problem with TensorFlow's data service when using the dynamic sharding policy, which is significantly slower compared to when the sharding policy is set to off. The user provides a simple example demonstrating the issue and seeks advice on what might be causing the slowdown.
- A user confirms the issue by running the provided code on Colab and experiencing the same performance problem.
- Number of comments: 11
-
GCS gfile operations fail in TF 2.17.0rc0 and 2.18 nightly when not running in GCP: This issue involves GCS operations using
tf.io.gfilefailing in TensorFlow versions 2.17.0rc0 and 2.18 nightly when not running on a Google Cloud Platform (GCP) virtual machine. The problem does not occur in TensorFlow 2.16 and is reproducible on various local environments, including macOS, Docker containers, GitHub actions, and Kaggle notebooks.- A user suggested that a specific commit might be influencing the issue, but testing is difficult due to the unavailability of older tf-nightly wheels on PyPI.
- Number of comments: 9
-
Training time for models: This issue is about a user experiencing significantly increased training times for their TensorFlow model after upgrading both their hardware and TensorFlow version, and seeking advice on whether this is expected behavior or if there are optimizations they can apply. The user also notes that while the error calculations during training have improved, the CPU utilization remains suboptimal, and they are considering upgrading to Keras 3 to potentially resolve these issues.
- The user received a suggestion to ensure their environment and TensorFlow version are optimized for the new hardware and to check if TensorFlow is utilizing all available cores.
- Number of comments: 7
-
[TFLite] TfLiteInterpreterInvoke has memory leak and increasing duration time: This issue reports a memory leak and increasing duration time when using the
TfLiteInterpreterInvokefunction in TensorFlow Lite's C API, observed across different TensorFlow versions and mobile platforms. The user has provided a code snippet and detailed observations, noting that the problem persists regardless of the TensorFlow version, device, or platform, and is not related to the use of NNAPI or GPU delegate options.
- The user confirmed that commenting out
TfLiteInterpreterInvokeprevents the issue, indicating the problem might be related to this function, and they are seeking further suggestions.- Number of comments: 7
1.3 Top 5 Quiet Issues:
We consider quiet issues to be issues that have been opened in this project for the longest time. The team should work together to get these issues resolved and closed as soon as possible.
-
error: no such file or directory: 'v2' when Bazel build for macOS: This issue describes a problem encountered while attempting to build TensorFlow C library on macOS using Bazel, where the build process fails with an error indicating that the file or directory 'v2' cannot be found. The user provides detailed steps and configurations used during the build process, including specific versions of TensorFlow, Bazel, and other tools, and highlights the error message received from the compiler.
- Open for 21 days, 04 hours, 01 minutes
-
GCS gfile operations fail in TF 2.17.0rc0 and 2.18 nightly when not running in GCP: This issue describes a bug in TensorFlow versions 2.17.0rc0 and 2.18 nightly where GCS operations using
tf.io.gfilefail when not running on a Google Cloud Platform (GCP) virtual machine. The problem causes the command to hang and eventually fail after 10 retries, with the error indicating issues in obtaining a Google authentication bearer token and resolving the host name.- Open for 21 days, 03 hours, 03 minutes
-
Immediate Assistance Required: Issue with Converting Keras Model to TFLite: This issue pertains to a problem encountered while converting a Keras model to TensorFlow Lite (TFLite), specifically an
AttributeErrorindicating that the 'Sequential' object lacks the attribute '_get_save_spec'. The error occurs during the model conversion process, preventing the successful export of a quantized TFLite model.- Open for 20 days, 18 hours, 10 minutes
-
Too many duplicate debug logs: This issue pertains to the presence of numerous duplicate debug logs when using TensorFlow Lite in an Android application. The user is seeking guidance on how to eliminate these redundant logs while utilizing the TensorFlow Lite library version 2.16.1.
- Open for 20 days, 10 hours, 28 minutes
-
Cannot take the length of shape with unknown rank.: This issue pertains to a ValueError encountered when attempting to train a U-Net model using TensorFlow, specifically arising from the inability to determine the length of a shape with an unknown rank. The error occurs during the execution of the
model.fitfunction, indicating a problem within thereduce_to_samplewise_valuesfunction in the Keras metrics module.- Open for 20 days, 06 hours, 01 minutes
1.4 Closed Issues
Closed Issues This Week: 13
Average Issue Close Time (This Week): 9.38 days
Average Issue Close Time (All Time): 5.56 days
Summarized Issues:
- TensorFlow 2.16 Crashes and Core Dumps: TensorFlow 2.16 on Linux Ubuntu 22.04.3 LTS has a bug where the
tf.raw_ops.CropAndResizeGradImageoperation crashes with a core dump due to tensor dimension mismatches. Additionally, specific inputs to thetf.raw_ops.UnsortedSegmentJoin,UnsortedSegmentMin,UnsortedSegmentMax,UnsortedSegmentProd, andUnsortedSegmentSumAPIs cause the session to crash by consuming all available RAM. Thetf.raw_ops.UnsortedSegmentJoinoperation also triggers an "Aborted (core dumped)" error.
- ONNX to TensorFlow Conversion Issues: Users encounter an
AttributeErrorwhen converting an ONNX model to a TensorFlow model using thetf2onnxlibrary. The error is due to the absence of thefrom_onnx_modelattribute in thetf2onnx.convertmodule. This issue highlights the need for better support and documentation for ONNX model conversion.
- Documentation for Custom Training Loops: There is a need for improved and more detailed documentation on creating custom training loops in TensorFlow. Users specifically request examples that address handling distributed training. This would help users better understand and implement custom training loops.
- Import Recognition in Visual Studio Code: Visual Studio Code fails to recognize the
tensorflow.kerasimport, despite TensorFlow being correctly installed. This issue affects users' ability to develop and debug TensorFlow applications within the IDE. It highlights the need for better integration and support for TensorFlow in development environments.
- UnicodeEncodeError on Windows 11: Users encounter a
UnicodeEncodeErrorwhen using thefitfunction in TensorFlow's Sequential model on Windows 11 with Python 3.12. The error is due to the system's default encoding not being set to UTF-8. This issue affects the usability of TensorFlow on Windows systems.
- TensorFlow to TensorFlow Lite Conversion Issues: Users face difficulties converting TensorFlow models to TensorFlow Lite, including potential problems with model accuracy and performance. Additional context and code are often needed to diagnose these issues effectively. This highlights the need for better tools and documentation for model conversion.
- Model Loading Error on Mac OS Sonoma: A bug in TensorFlow version 2.15.1 and Keras version 2.15.0 on Mac OS Sonoma prevents loading a saved model using
tf.keras.models.load_model. The error is due to the inability to import thekeras.src.engine.functionalmodule. This issue affects the usability of TensorFlow on Mac OS.
- ReLU Activation Function in TensorFlow Lite: Unexpected negative values appear in the output of a ReLU activation function within a TensorFlow Lite quantized model. This was later clarified to be a result of the quantization process and not an actual bug. It highlights the need for better understanding and documentation of quantization effects.
- Compatibility Issues with Keras 3.0: Errors occur while running a specific TensorFlow notebook due to compatibility changes in Keras 3.0. Users request an update to the notebook to resolve these errors. This issue highlights the need for maintaining compatibility and updating resources with new library versions.
- Segmentation Fault on Raspberry Pi 4: A segmentation fault occurs when running a TensorFlow Lite model in C++ on a Raspberry Pi 4. Users face difficulties converting the model to TFLite and running it natively, despite it working correctly with the TFLite interpreter in Python. This issue affects the deployment of TensorFlow Lite models on embedded systems.
- Coral USB Device Inference Issue: TensorFlow Lite initializes a Coral USB device on a Raspberry Pi 4 but runs inference on the CPU instead of the Coral device. Logs and CPU utilization measurements confirm this issue. It highlights the need for better support and troubleshooting for hardware acceleration.
- Model Loading Error on Ubuntu 22.0: A model saved using TensorFlow 2.15 on Kaggle cannot be loaded on an Ubuntu 22.0 system with TensorFlow 2.15. The error is a
TypeErrorrelated to deserializing the 'Functional' class from the Keras module. This issue affects the portability and usability of TensorFlow models across different environments.
1.5 Issue Discussion Insights
This section will analyze the tone and sentiment of discussions within this project's open issues within the past week to identify potentially heated exchanges and to maintain a constructive project environment.
- 👉🏻 Hackear Instagram: Um Guia Atualizado
- Toxicity Score: 0.70 (Inappropriate content, potential for negative reactions, promotional tone)
- This GitHub conversation involves a user named Artesanales who opened an issue with a misleading title and body content that promotes hacking services. The tone of the initial post is promotional and inappropriate for the platform, which is likely to trigger negative reactions from the community. There are no comments yet, but the nature of the post suggests potential for future tension and negative sentiment if the issue is not promptly addressed by moderators.
II. Pull Requests
2.1 Open Pull Requests
Open Pull Requests This Week: 3
Pull Requests:
- Broadcasting unequal rank inputs in TFL_BatchMatMulOp: This topic covers the addition of support for broadcasting unequal rank inputs in the
TFL_BatchMatMulOpwhen lowering totosa.matmulin TensorFlow. The pull request ensures that the operation can handle inputs of different ranks, which is crucial for various machine learning models. This enhancement improves the flexibility and robustness of the TensorFlow framework.
- KubernetesClusterResolver environment specification: This topic addresses the ability to specify whether the
KubernetesClusterResolveris used within a Kubernetes cluster or an external environment. The pull request resolves issue #70581 by adding this functionality, which is essential for users running TensorFlow in diverse deployment scenarios. This improvement helps in better configuration and management of TensorFlow clusters.
- Resolving build error in TensorFlow Lite label_image example: This topic focuses on fixing the build error in the TensorFlow Lite label_image example when using CMake. The pull request addresses issue #70659 and ensures that the example can be built successfully, which is important for developers using TensorFlow Lite. This fix enhances the usability and reliability of TensorFlow Lite examples.
2.2 Closed Pull Requests
Summarized Pull Requests:
As of our latest update, there are no closed pull requests for the project this week.
2.3 Pull Request Discussion Insights
This section will analyze the tone and sentiment of discussions within this project's open pull requests within the past week to identify potentially heated exchanges and to maintain a constructive project environment.
Based on our analysis, there are no instances of toxic discussions in the project's open pull requests from the past week.
III. Commits
3.1 Commits
Commits This Week: 101
Summarized Commits:
- Batch Stats Module Enhancement: This commit introduces a feature to the batch stats module that adds a cumulative query count for each model, enhancing the tracking and analysis of model queries over time.
- XLA:Python Optimization: This commit introduces a minor optimization in the XLA:Python project by defining a
tp_hashslot instead of a__hash__method to improve performance slightly. It also notes thattp_richcompareis marginally faster to dispatch than__eq__, based on profiling observations.
- Redundant Code Removal: This commit removes a redundant bounds check in the code, as the condition is already verified in the EvaluateEpilogue function where
thread_id[1]is constrained to {0, 0}. Another commit removes the dependency onlibtensorflow_framework.so.2, which was initially added as a workaround for a Bazel bug that now appears to be resolved.
- Compilation and Build Fixes: Several commits address and fix compilation errors and build failures in various tests and targets within the XLA service on GPU, ensuring compatibility and successful execution in the open-source setting.
- Dependency Management: This commit inlines the XLA_DEVICE_DEPS and runs build_cleaner to avoid the antipattern of factoring out common dependencies, which complicates dependency management and can lead to the inclusion of unnecessary dependencies in targets that reference the "common" dependencies.
- Custom Call Registry API: This commit introduces the recording of command buffer custom call targets for the legacy custom call registry API in XLA:GPU, providing temporary support for lowering TE custom call kernels to command buffers via a whitelist until the FFI API migration is complete.
- Performance Improvements: This commit updates the
ConstraintExpressionto utilizellvm::SmallVectorinternally, resulting in a 20% improvement in compile-time performance on benchmarks. Another commit optimizes the computation of buffer sizes in bytes by using statically known buffer ranks, resulting in performance improvements for various benchmarks.
- Forward Compatibility and Version Updates: Multiple commits update the forward compatibility horizon and GraphDef version to ensure the project remains compatible with future changes and developments.
- Automated Code Changes: Several commits involve automated code changes, as indicated by the commit messages and associated PiperOrigin-RevIds.
- Refactoring and Code Cleanup: Various commits involve refactoring code, such as splitting multi-row reduction functionality into a separate class, renaming functions for clarity, and removing redundant include directives to streamline the codebase.
- New Features and Enhancements: This commit introduces support for uint8_t dot operation tests and integrates the necessary functionality into the HLO evaluator. Another commit introduces broadcast splitting before sharding propagation in XLA:SPMD to prevent unwanted sharding propagation.
- Documentation Updates: This commit updates the
docs/operation_semantics.mddocument by adding detailed descriptions for several element-wise binary operations. Another commit adds documentation and examples on how to use themultihost_hlo_runner.
- Bug Fixes and Error Handling: Several commits address and fix bugs, such as correcting file paths in target patterns, fixing constructor conversion errors, and enhancing error messages for better clarity.
- LLVM Integration: Multiple commits integrate the LLVM project at specific commits, updating the project's LLVM usage to align with the changes made in those particular revisions.
- Test Enhancements: This commit introduces new test cases and benchmarks to verify the deduplication of HLO operations and the management of execution threads in the XLA compiler.
- Build Configuration Updates: This commit updates the build configuration to use specific CPU tag filters for the CPU build and CUDA tag filters for the CUDA build, ensuring appropriate build settings for different environments.
- FFI and PJRT Enhancements: This commit introduces FFI custom call thunk runtime support to the PJRT CPU client and includes a benchmark to evaluate the PJRT CPU Client's performance.
- Miscellaneous Updates: Various commits include updates such as adding a defensive mechanism for exception handling in user FFI calls, updating the google/fuzztest dependency, and enhancing the
RemapPlan::DebugString()function.
IV. Contributors
4.1 Contributors
Active Contributors:
We consider an active contributor in this project to be any contributor who has made at least 1 commit, opened at least 1 issue, or created at least 1 pull request in the past month.
| Contributor | Commits | Pull Requests | Issues |
|---|---|---|---|
| A. Unique TensorFlower | 23 | 0 | 0 |
| Eugene Zhulenev | 15 | 0 | 0 |
| Alexander Belyaev | 6 | 0 | 0 |
| Adrian Kuegel | 6 | 0 | 0 |
| TensorFlower Gardener | 5 | 0 | 0 |
| Johannes Reifferscheid | 5 | 0 | 0 |
| Henning Becker | 3 | 0 | 0 |
| Oleg Shyshkov | 3 | 0 | 0 |
| David Dunleavy | 3 | 0 | 0 |
| Leo-Lifeblood | 0 | 0 | 3 |
| damjandakic93 | 0 | 0 | 3 |
| Benjamin Chetioui | 2 | 0 | 0 |
| Dirk Hornung | 2 | 0 | 0 |
| Dimitar (Mitko) Asenov | 2 | 0 | 0 |
| akhilgoe | 2 | 0 | 0 |
| David Majnemer | 2 | 0 | 0 |
| sanowl | 0 | 2 | 0 |
| msteiner-google | 0 | 1 | 1 |
| NobuoTsukamoto | 0 | 1 | 1 |
| x0w3n | 0 | 0 | 2 |
| lgeiger | 0 | 0 | 2 |
| orikkwak | 0 | 0 | 2 |
| MengXiangDing | 0 | 0 | 2 |
| fujunwei | 0 | 0 | 2 |
| ravin00 | 0 | 0 | 2 |
| Paweł Paruzel | 1 | 0 | 0 |
| Greg Olechwierowicz | 1 | 0 | 0 |
| Alan Kelly | 1 | 0 | 0 |
| Shan Han | 1 | 0 | 0 |
| Ilia Sergachev | 1 | 0 | 0 |
| Juhyun Lee | 1 | 0 | 0 |
| Olli Lupton | 1 | 0 | 0 |
| Kanvi Khanna | 1 | 0 | 0 |
| Sergey Kozub | 1 | 0 | 0 |
| Jaroslav Sevcik | 1 | 0 | 0 |
| Tori Baker | 1 | 0 | 0 |
| Aiden Grossman | 1 | 0 | 0 |
| Tongfei Guo | 1 | 0 | 0 |
| Matt Callanan | 1 | 0 | 0 |
| Abhinav Gunjal | 1 | 0 | 0 |
| Anlun Xu | 1 | 0 | 0 |
| Gregory Pataky | 1 | 0 | 0 |
| Sandeep Dasgupta | 1 | 0 | 0 |
| Clive Verghese | 1 | 0 | 0 |
| pizzud | 1 | 0 | 0 |
| Aliia Khasanova | 1 | 0 | 0 |
| Ce Zheng | 1 | 0 | 0 |
| kavilivishnu | 0 | 1 | 0 |
| cboss6 | 0 | 1 | 0 |
| WilliamRoyNelson | 0 | 1 | 0 |
| LakshmiKalaKadali | 0 | 1 | 0 |
| mattbahr | 0 | 1 | 0 |
| hsharsha | 0 | 1 | 0 |
| belitskiy | 0 | 1 | 0 |
| Mirgahney | 0 | 1 | 0 |
| psunn | 0 | 1 | 0 |
| dingjieliu | 0 | 0 | 1 |
| arthurflor23 | 0 | 0 | 1 |
| Barooni-saeede | 0 | 0 | 1 |
| jsasswis | 0 | 0 | 1 |
| captainst | 0 | 0 | 1 |
| AnyLifeZLB | 0 | 0 | 1 |
| MAlmuzaini | 0 | 0 | 1 |
| SMH17 | 0 | 0 | 1 |
| Dustglaxy | 0 | 0 | 1 |
| Jerry-yz | 0 | 0 | 1 |
| GreatKampfstern | 0 | 0 | 1 |
| gachombamuthiri | 0 | 0 | 1 |
| johnlarkin1 | 0 | 0 | 1 |
| willendzw | 0 | 0 | 1 |
| ninjaguardian | 0 | 0 | 1 |
| MatthiasNeumueller | 0 | 0 | 1 |
| chiamp | 0 | 0 | 1 |
| myh1234567 | 0 | 0 | 1 |
| hui-li-xf | 0 | 0 | 1 |
| kindaTall | 0 | 0 | 1 |
| BenCrulis | 0 | 0 | 1 |
| H4dr1en | 0 | 0 | 1 |
| maxwellbrien1990 | 0 | 0 | 1 |
| Nayana-ibm | 0 | 0 | 1 |
| yanncalec | 0 | 0 | 1 |
| Suryansh-patidar | 0 | 0 | 1 |
| kupix | 0 | 0 | 1 |
| feranick | 0 | 0 | 1 |
| cpappasILMX | 0 | 0 | 1 |
| ThorvaldAagaard | 0 | 0 | 1 |
| cz2h | 0 | 0 | 1 |
| yousifj129 | 0 | 0 | 1 |
| manswad | 0 | 0 | 1 |
| SsomsakTH | 0 | 0 | 1 |
| weebao | 0 | 0 | 1 |
| basi880830 | 0 | 0 | 1 |
| WeichenXu123 | 0 | 0 | 1 |
| NoranNabil | 0 | 0 | 1 |
| mrezaanony | 0 | 0 | 1 |
| kkk-jpeg | 0 | 0 | 1 |
| CV-42 | 0 | 0 | 1 |
| tomneo2004 | 0 | 0 | 1 |
| pekaalto | 0 | 0 | 1 |
| npanpaliya | 0 | 0 | 1 |
| pataneanant | 0 | 0 | 1 |
| suyash-narain | 0 | 0 | 1 |
| hamzaziizzz | 0 | 0 | 1 |
| arpanptnk85 | 0 | 0 | 1 |
| luanft | 0 | 0 | 1 |
| KarenMars | 0 | 0 | 1 |
| Sklaebe | 0 | 0 | 1 |
| Aweptimum | 0 | 0 | 1 |
| alykhantejani | 0 | 0 | 1 |
| baqwas | 0 | 0 | 1 |
| anilkumarvvss | 0 | 0 | 1 |
| asp616848 | 0 | 0 | 1 |
| hiwothadush | 0 | 0 | 1 |
| junwha0511 | 0 | 0 | 1 |
| ramonhollands | 0 | 0 | 1 |
| immusferr | 0 | 0 | 1 |
| syedhamzamohiuddin | 0 | 0 | 1 |
| beatsea20 | 0 | 0 | 1 |
| panhu | 0 | 0 | 1 |
| tirk999 | 0 | 0 | 1 |
| Non1187 | 0 | 0 | 1 |
| Sachi-27 | 0 | 0 | 1 |
| bossebandowski | 0 | 0 | 1 |
| SakshiFadnavis2003 | 0 | 0 | 1 |
| RaulCastillo547 | 0 | 0 | 1 |
| dwang6524 | 0 | 0 | 1 |
| kiriti-pendyala | 0 | 0 | 1 |
| durgas4 | 0 | 0 | 1 |
| devapriyas2001 | 0 | 0 | 1 |
| buttaRahul | 0 | 0 | 1 |
| WojciechRynczuk | 0 | 0 | 1 |
| hotamago | 0 | 0 | 1 |
| AdityaB-01 | 0 | 0 | 1 |
| babvijayb | 0 | 0 | 1 |
| Artesanales | 0 | 0 | 1 |
| yayale1 | 0 | 0 | 1 |