Weekly GitHub Report for Tensorflow: April 07, 2025 - April 14, 2025 (14:17:45)
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.
Table of Contents
I. News
1.1 Recent Version Releases:
The current version of this repository is v2.19.0
1.2 Version Information:
The TensorFlow 2.19.0 release, created on March 5, 2025, introduces breaking changes to the LiteRT C++ and Python APIs, including the transition of tf.lite.Interpreter to a new location with a deprecation warning, and adds support for bfloat16 in the tfl.Cast operation. Additionally, the release discontinues publishing libtensorflow packages, though they remain accessible via PyPI, and features contributions from a diverse group of developers.
II. Issues
2.1 Top 5 Active Issues:
We consider active issues to be issues that that have been commented on most frequently within the last week. Bot comments are omitted.
-
Convolution FP32 in oneDNN Changed from gemm:acl to gemm:ref in TensorFlow 2.16: This issue highlights a performance degradation in TensorFlow version 2.16 and later, where the convolution operation for FP32 data type in oneDNN was changed from using gemm:acl to gemm:ref, resulting in slower performance compared to version 2.15. The user provides detailed steps and a script to reproduce the issue, requesting confirmation on whether this change was intentional and seeking a fix or workaround to restore the previous performance levels.
- The comments involve a request for a code snippet to reproduce the issue, which the user provides along with detailed steps and expected versus observed behavior. The user also requests confirmation on the intentionality of the change and seeks input from multiple contributors for further investigation.
- Number of comments this week: 3
-
(deprecated arguments) (deprecated arguments) (deprecated arguments): This issue is about a documentation bug in TensorFlow's API documentation, specifically related to deprecated arguments in the
tf.functionfeature. The user has confirmed that the bug is reproducible with TensorFlow Nightly and has provided a link to the relevant documentation page where the issue is observed.- The comments discuss the appropriate repository for reporting the issue, with a suggestion to open it against the documentation repository. However, it is noted that the suggested repository does not have an "issues" tab, leading to some confusion and a lighthearted acknowledgment of the oversight.
- Number of comments this week: 3
-
No TPU platform registeredwith tf.estimator and TPUv5: This issue involves a bug encountered when using TensorFlow'stf.estimatorwith a TPUv5, where the code that works on TPUv3 fails with a segmentation fault and an error message indicating "No TPU platform registered." The user is attempting to run training code on a TPUv5p via Google Kubernetes Engine (GKE) and has provided a detailed setup and code snippet, but is unsure if the problem is due to a compatibility issue or a mistake in their setup.- The comments discuss attempts to reproduce the issue using an existing MNIST example on TPU, which also fails with a similar error message about the TPU platform not being registered. The user provides a detailed script and mentions that the error slightly differs when run on a TPUv3 VM, attaching the complete error output for further analysis.
- Number of comments this week: 2
-
'OK' is not a member of 'tsl::Status': This issue is about a bug encountered in a TensorFlow project where the 'OK' status is not recognized as a member of 'tsl::Status', leading to compilation errors in the GPUInstrumentOp.cc file. The problem persists even when using TensorFlow version 2.14, and the user has attempted to resolve it by following a guide for creating operations, but the solution only works for CPU, not GPU.
- The comments suggest that TensorFlow 2.14 is no longer supported, and the user is advised to try a newer version. The user responds that the issue persists even with TensorFlow 2.17, indicating that the problem is not resolved by simply updating the TensorFlow version.
- Number of comments this week: 2
Since there were fewer than 5 open issues, all of the open issues have been listed above.
2.2 Top 5 Stale Issues:
We consider stale issues to be issues that has had no activity within the last 30 days. The team should work together to get these issues resolved and closed as soon as possible.
- TF-TRT Warning: Could not find TensorRT: This issue involves a user experiencing difficulties with TensorFlow on an Ubuntu 22.04 system, specifically encountering a warning that TensorRT could not be found despite multiple attempts to install and configure the software within an Anaconda environment. The user suspects that the problem may be related to driver compatibility issues with their NVIDIA RTX 3050 TI GPU, as they are forced to use an older driver version (535) instead of the automatically installed 550 version, which does not work on their multi-monitor setup.
SystemErrorintf.ensure_shapeandtf.compat.v1.ensure_shapewhendtypeofshapeistf.uint64and its value is too large.: This issue pertains to a bug in TensorFlow where usingtf.ensure_shapeortf.compat.v1.ensure_shapewith ashapeofdtypetf.uint64and a value close to 2^64 results in aSystemErrorandOverflowError. The problem has been reproduced with TensorFlow Nightly on a Linux Ubuntu 20.04 system using Python 3.10, and it specifically occurs when the shape is set to a large value, such as[18446743219011059112, 1], in eager execution mode.- Feature Request: Integrate different Digital Signal Processing into tf.signal: This issue is a feature request to integrate advanced Digital Signal Processing (DSP) functionalities into TensorFlow's
tf.signalmodule, similar to those available in the PyTorch ecosystem, particularly from the julius library. The integration aims to provide native tools for complex audio data augmentation, enhancing the capabilities for researchers and developers working on audio-related projects by streamlining workflows and reducing dependency on external libraries. - [DOCS] Missing complex input for Round op: This issue highlights a documentation bug in TensorFlow's
Roundoperation, where the official documentation incorrectly states that a complex tensor can be used as input, but in practice, this does not work as expected. The user reports that they must apply theRoundoperation separately to the real and imaginary parts of the tensor to achieve the desired result, indicating a discrepancy between the documented and actual behavior of the function. - tf.raw_ops.Unbatch aborts with "Check failed: d < dims()": This issue involves a bug in TensorFlow version 2.17 where the
tf.raw_ops.Unbatchoperation aborts unexpectedly with an error message indicating a failed check on tensor dimensions. The problem has been reproduced using TensorFlow Nightly on a Linux Ubuntu 20.04.3 LTS system with Python 3.11.8, and the error occurs when executing a specific standalone code snippet provided in the issue description.
2.3 Open Issues
This section lists, groups, and then summarizes issues that were created within the last week in the repository.
Issues Opened This Week: 11
Summarized Issues:
- Performance Issues in TensorFlow Operations: There are performance-related problems in TensorFlow where unexpected data type casting occurs during operations, specifically from int32 to float64, despite no explicit casting in the source code. This issue has been reproduced using TensorFlow Nightly, indicating a persistent problem across versions.
- Conversion and Compatibility Issues with TensorFlow.js: Users face a
ValueErrordue to unsupported operations likeErfcwhen converting TensorFlow models to TensorFlow.js format usingtensorflowjs_converter. The issue persists even when avoiding theskip_op_checkconfiguration, highlighting a gap in compatibility between TensorFlow and TensorFlow.js.
- Denial of Service Vulnerabilities in TensorFlow: TensorFlow version 2.19.0 has bugs in operations like
tf.raw_ops.DenseCountSparseOutputandtf.raw_ops.BlockLSTMGradV2, causing "Aborted (core dumped)" errors. These errors can potentially be exploited for denial of service attacks, as demonstrated by code examples using empty tensors.
- Documentation and Compilation Bugs in TensorFlow: There are documentation bugs in the TensorFlow API regarding deprecated arguments in function compilation features. Additionally, compilation issues arise in
GPUInstrumentOp.ccdue to unrecognized members, affecting users even when following official guides.
- TensorFlow Model Output Corruption with
tf.pad: A bug in TensorFlow version 2.15.1 causes corrupted model outputs when usingtf.padwithmode='reflect'on large input arrays. This issue does not occur withnp.pad, indicating a specific problem with TensorFlow's padding implementation.
- Installation Issues with TensorFlow on macOS: The
.whlfile built from TensorFlow source on macOS lacks the specific system version in its filename, leading to installation errors. Users must manually add the version number to avoid the file being reported as unsupported on the platform.
- Import Errors with TensorFlow on Windows: Users encounter ImportErrors related to DLL load failures when importing TensorFlow on Windows 11, affecting versions 2.13.0 and 2.19.0. These errors persist even with TensorFlow Nightly, indicating a broader issue with the native TensorFlow runtime on Windows.
2.4 Closed Issues
This section lists, groups, and then summarizes issues that were closed within the last week in the repository. This section also links the associated pull requests if applicable.
Issues Closed This Week: 15
Summarized Issues:
- Abortion Errors in TensorFlow 2.20.0-dev20250302: Several issues in TensorFlow version 2.20.0-dev20250302 involve abortion errors that can potentially be exploited for denial of service attacks. These errors occur in various functions such as
tensorflow.keras.remat,tf.conv, andtf.compat.v2.nn.depthwise_conv2d_backprop_input, often triggered by invalid indices or parameter values.
- ImportError Due to DLL Issues on Windows: Multiple issues report ImportErrors on Windows platforms due to failures in loading TensorFlow's native runtime, often related to DLL initialization failures. These problems occur across different TensorFlow versions and Python environments, highlighting compatibility challenges and the need for proper environment setup.
- Build and Installation Issues on Various Platforms: Users face challenges building or installing TensorFlow on different platforms, such as using NVCC and Clang on Alma 9.5 or installing on a MacBook Air 2018. These issues often stem from compatibility problems with compilers or package managers, requiring specific configurations or updates.
- Compatibility with Newer GPU Models and NumPy Versions: There are concerns about TensorFlow's compatibility with newer GPU models and NumPy versions. Users report difficulties with TensorFlow not working on updated GPUs and suggest updating the NumPy version constraint to prevent errors with NumPy 2.x.
2.5 Issue Discussion Insights
This section will analyze the tone and sentiment of discussions within this project's open and closed issues that occurred within the past week. It aims 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 or closed issues from the past week.
III. Pull Requests
3.1 Open Pull Requests
This section provides a summary of pull requests that were opened in the repository over the past week. The top three pull requests with the highest number of commits are highlighted as 'key' pull requests. Other pull requests are grouped based on similar characteristics for easier analysis. Up to 25 pull requests are displayed in this section, while any remaining pull requests beyond this limit are omitted for brevity.
Pull Requests Opened This Week: 3
Key Open Pull Requests
1. Fix typos in documentation strings: This pull request addresses the correction of typographical errors in the documentation strings of the TensorFlow project, as indicated by the commit message and the detailed description provided by the contributor, who requests the necessary actions to be taken.
- URL: pull/91179
- Merged: No
- Associated Commits: 1e9dc
2. Enable Android XNNPACK delegate num_threads with Interpreter.Options.…: This pull request addresses the issue of the setNumThreads() API not functioning correctly on Android by defaulting to a single thread, which negatively impacts performance, and proposes using xnnpack_create_threadpool with context to properly create a thread pool, as detailed in the commit by Yi Jin Sun.
- URL: pull/91181
- Merged: No
- Associated Commits: 81875
3. Added hello_contributors.txt as my first contribution: This pull request involves the addition of a file named "hello_contributors.txt" to the TensorFlow project as the contributor's first submission, indicating their intention to begin learning and contributing more to the project.
- URL: pull/91241
- Merged: No
- Associated Commits: dd311
3.2 Closed Pull Requests
This section provides a summary of pull requests that were closed in the repository over the past week. The top three pull requests with the highest number of commits are highlighted as 'key' pull requests. Other pull requests are grouped based on similar characteristics for easier analysis. Up to 25 pull requests are displayed in this section, while any remaining pull requests beyond this limit are omitted for brevity.
Pull Requests Closed This Week: 11
Key Closed Pull Requests
1. Exported pr 741299485: This pull request involves multiple updates and improvements to the TensorFlow project, including extracting TMA metadata from the Triton module, activating additional tests in fusion_emitter_device_legacy_port_test.cc, implementing a pass to split the GEMM K dimension for better compute utilization, updating and enabling legacy matmul emitter tests, adding an OnDestroy callback to Collectives, implementing replica group deduplication for HloReplicationAnalysis, instrumenting batching and queueing delay metrics, allowing multihost runner to use entry computation layouts, and various other enhancements and fixes across the codebase, although it was ultimately not merged.
- URL: pull/91239
- Merged: No
- Associated Commits: 4a33c, be098, f4288, 84226, 9e138, 0b72d, e3546, fa854, 24ef4, 9aabd, 2a53a, 37c09, f161b, 1a3aa, 9f47a, 45b5e, df2fb, e5c78, 9893e, 97da0, c31ed, 93457, 6977a, 649a4, 644d8, 2beac, 9b091, 77efc, d0f75, 052fe, abf78, 964bd, 23b9e, 94db8, a172d, 0ff0b, 34b1a, c0093, 3f0e0, 884af, aa6fc, 5ed40, a3689, c96bd, e6e9c, 51b69, 840ac, 01ee4, 96b4c, 74959, aed1f, 2a994, 18054, 7506b, 1a584, 6e29b, 880c6, fa367, b549e, b18b4, ae30a, c588a, a7dbd, 22885, 94a3e, 7dc0b, 270f6, 6ffe8, 957ad, 53c9e, d9a88, 3cf28, 2df2b, 8a4aa, 2b9a5, 4934f, e8bbb, b875c, 6efef, 84719, bb9d4, 02d7d, cdfa5, 5cba8, f2b3e, 7230b, c2636, 62fb2, 43312, 72a9a, f505a, 16186, 62afe, 8ec51, 37946, f7624, 13bf1, 73d9b, a7d97, 101c8, a804e, 7ea58, 85071, 3b77c, 92ea2, fc08d, c300b, 8f343, a9fb0, 3ff1a, 6bc18, cb94d, d38b0, 4fcda, 6d473, ed315, 6358e, 63db9, 0e2c3, 43d5e, 11e1b, 34d6e, dbaf7, bc491, 3b1f4, d59b8, 5f309, c5d30, fe99c, 09dfb, 9697f, 33a7c, 5f31a, c995c, 852a5, e27d9, 33c50, d4b08, c4a23, c060c, 97c65, 1d579, 4a9af, 67878, 0602a, 3c7b3, 8fc42, 02819, 90d83, dcb4d, 87d8e, 24023, 09571, ba503, 113cc, ded04, 9686f, 82285, e7056, aac64, 01285, 458f4, 99f88, 41ca7, 3949b, 3fc00, a18e4, 0377f, baf70, c9ff6, 90e43, 9454e, bbc16, 42a3a, 0e35d, 2f4e9, 46897, 46e7e, 12feb, 25e9a, 9a463, 47996, 4ac19, 5c213, d16a8, dc4fa, 48173, dbf18, 76e3f, cd9d1, f99f0, ad147, 85b53, 0b585, b83ce, 3c3bc, 0fc2b, d3e4b, 55e97, 2aff0, 7f2a0, 82638, 6c5b9, 467c2, 9ffc0, 93f72, 464bc, f5a40, 71e33, cc430, 5f5b7, a59e9, ba906, a578a, 624a7, 3fbe7, 99e7a, 57072, 72827, 363a0, c211d, 50d47, a2f96, f76cd, 39097, f0280, 7c589, bdbb0, 4e883, 5f5db, c552b, a1d3e, 472a5, f803a, 7e060, 05d5c, c64f2, 18090, 8f8f3, 3b18a, 3fac2, 3be27, 1ae55, 1ba18, 9d769, ebd9a, 6a8c8, b61a8, fb5c1, 9f5db
2. Fix compilation error due to overloads of cub::ThreadLoadVolatilePointer: This pull request addresses a compilation error in the TensorFlow project by applying a fix to the gpu_prim.h file, specifically targeting the overloads of cub::ThreadLoadVolatilePointer, to resolve issues encountered when compiling sparse_grad_op_gpu.cu.cc with the clang compiler, as detailed in a previous commit.
- URL: pull/90494
- Merged: 2025-04-08T19:11:15Z
3. Remove ambiguous inherited constructor in default_quant_params.cc: This pull request addresses an issue in the TensorFlow project by removing an ambiguous inherited constructor in the default_quant_params.cc file, which was causing complaints from the GCC compiler, and provides a trivial and harmless fix that resolves issue #84977.
- URL: pull/90558
- Merged: 2025-04-08T11:48:34Z
Other Closed Pull Requests
- MLIR TOSA Legalization and Fixes: This topic covers several pull requests that address issues and enhancements in the MLIR TOSA component of TensorFlow. The pull requests include the legalization of the LOG operator for int8 and int16 data types, resolving issues with failing lit tests, and ensuring correct maxpool2d operator shape through input slicing.
- Documentation and Typographical Corrections: This pull request focuses on correcting typographical errors in the documentation strings of the TensorFlow project. The changes were successfully merged, improving the overall quality and readability of the documentation.
- Build and Compiler Fixes: This pull request resolves a build failure issue when using NVCC with Clang by correcting the misidentification of the compiler in
gpu_device_functions.h. The fix ensures that the correct functions are used, preventing build errors related to the__clang__definition.
- Test Enhancements and Bug Fixes: This topic includes pull requests that enhance test cases and fix bugs in the TensorFlow project. The changes involve fixing the
transpose_conv2dlit test and adding a test case for verifying the concatenation of quantized operands and results in the TOSA MLIR framework.
3.3 Pull Request Discussion Insights
This section will analyze the tone and sentiment of discussions within this project's open and closed pull requests that occurred within the past week. It aims 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 or closed pull requests from the past week.
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, created at least 1 pull request, or made more than 2 comments in the last month.
If there are more than 10 active contributors, the list is truncated to the top 10 based on contribution metrics for better clarity.
| Contributor | Commits | Pull Requests | Issues | Comments |
|---|---|---|---|---|
| tensorflower-gardener | 61 | 0 | 0 | 0 |
| Venkat6871 | 5 | 3 | 0 | 24 |
| mihaimaruseac | 1 | 0 | 0 | 28 |
| maludwig | 1 | 0 | 0 | 16 |
| bchetioui | 13 | 0 | 0 | 0 |
| ezhulenev | 13 | 0 | 0 | 0 |
| default1360 | 0 | 0 | 11 | 2 |
| kossyrev-bg | 0 | 0 | 1 | 10 |
| lhutton1 | 6 | 4 | 0 | 0 |
| plopresti | 3 | 3 | 3 | 1 |
Access Last Week's Newsletter: