Weekly GitHub Report for Tensorflow: April 28, 2025 - May 05, 2025 (12:00:02)

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
  • 1.2. Other Noteworthy Updates
• II. Issues
  • 2.1. Top 5 Active Issues
  • 2.2. Top 5 Stale Issues
  • 2.3. Open Issues
  • 2.4. Closed Issues
  • 2.5. Issue Discussion Insights

• III. Pull Requests
  • 3.1. Open Pull Requests
  • 3.2. Closed Pull Requests
  • 3.3. Pull Request Discussion Insights

• IV. Contributors
  • 4.1. Contributors

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 in 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.

1. Tensorflow for Arch Linux: This issue is a feature request for TensorFlow to support Arch Linux, as the user is experiencing dependency issues due to Arch's rolling release nature while working on a project using TensorFlow. The user has been using Linux for its lightweight nature compared to Windows and is encountering difficulties with TensorFlow on Arch, which they hope can be resolved with official support.

   • The comments discuss the lack of official support for Arch Linux by the TensorFlow team, with suggestions to use pip or conda for installation. The user explains that these methods lead to version conflicts in their dedicated environment. Alternatives like hatch or Nix dev-shells are suggested for better environment management.
   • Number of comments this week: 5

2. Compiling 2.19.0 through cmake on Windows, there are too many exported DLL symbols and an increase in disk size: This issue involves a problem with compiling TensorFlow version 2.19.0 using CMake on Windows, where there is an unusually high number of exported DLL symbols, leading to a larger disk size compared to version 2.18.0. The user is seeking a solution to control the number of exported symbols and reduce the DLL size, as they believe the current behavior is abnormal.

   • The comments suggest that the issue may be due to a lack of control over Windows DLL export symbols, with additional templates and internal classes being exported without filtering in version 2.19.0. A proposed solution is to use the CMake flag -DCMAKE_WINDOWS_EXPORT_ALL_SYMBOLS=OFF and introduce proper export control macros or a manual .def file to manage public and internal symbols, which the user acknowledges as feasible but hopes for an official fix.
   • Number of comments this week: 2

3. No way to use tensorflow with cuda on windows. cudaGetDevice() failed.: This issue involves a user experiencing difficulties using TensorFlow with CUDA on Windows, specifically encountering a "cudaGetDevice() failed" error. The user initially resolved the error by reinstalling TensorFlow-GPU and setting the environment variable CUDA_VISIBLE_DEVICES, but remains uncertain about the actual usage of the GPU during execution.

   • The user fixed the initial error by reinstalling TensorFlow-GPU and setting the CUDA_VISIBLE_DEVICES environment variable, which allowed GPU detection. However, they are unsure if the GPU is being used during execution and seek confirmation. A response suggests upgrading to a newer TensorFlow version for better GPU support and offers assistance with the upgrade process.
   • 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.

1. SystemError in tf.ensure_shape and tf.compat.v1.ensure_shape when dtype of shape is tf.uint64 and its value is too large.: This issue pertains to a bug in TensorFlow where using tf.ensure_shape or tf.compat.v1.ensure_shape with a shape of dtype tf.uint64 and a value close to 2^64 results in a SystemError and OverflowError. The problem is reproducible in TensorFlow version 2.15 on a Linux Ubuntu 20.04 system, and it occurs when the shape value is excessively large, as demonstrated with the example shape = tf.constant([18446743219011059112, 1], dtype=tf.uint64).
2. 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.signal module, similar to those available in the PyTorch ecosystem, particularly the julius library. The integration aims to enhance TensorFlow's capabilities in audio data augmentation, providing researchers and developers with native tools for complex audio processing, thereby improving workflow efficiency and model generalization without relying on external libraries.
3. [DOCS] Missing complex input for Round op: This issue highlights a documentation bug in TensorFlow's Round operation, 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 the Round operation separately to the real and imaginary parts of the tensor to achieve the desired result, indicating a discrepancy between the documentation and the actual functionality.
4. 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.Unbatch operation aborts unexpectedly with an error message "Check failed: d < dims()". The problem occurs on a Linux Ubuntu 20.04.3 LTS system using Python 3.11.8, and it has been reproduced with TensorFlow Nightly, indicating a persistent issue in the software. Since there were fewer than 5 open issues, all of the open issues have been listed above.

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: 10

Summarized Issues:

• Compilation Issues on Windows and ARM Platforms: Compiling TensorFlow version 2.19.0 on Windows using CMake results in an abnormal increase in exported DLL symbols and disk size due to uncontrolled symbol exports. Additionally, compiling with MKL support on an ARM SVE platform using Clang 17.0.6 fails due to a missing csr member in dnnl::memory::desc.
  • issues/92275, issues/92279

• GPU and CUDA Compatibility Issues: Users face challenges with TensorFlow's GPU utilization and CUDA compatibility, including illegal GPU memory access errors on Ubuntu 20.04 with CUDA 12.5.1 and uncertainty about GPU usage on Windows despite detection. These issues highlight the need for better GPU integration and error handling.
  • issues/92413, issues/92523

• TensorFlow Lite and Hexagon Delegate Issues: Users are unable to access the oemconfig.so file necessary for the Hexagon delegate on Snapdragon devices, despite following guides and trying various methods. This results in dynamic loading failures, indicating a need for improved support and documentation for TensorFlow Lite on specific hardware.
  • issues/92488

• Performance Discrepancies on Apple Silicon: TensorFlow versions 2.12, 2.16.2, and 2.19 show lower accuracy and slower performance on M4 MacBooks with Apple Silicon compared to Google Colab's GPU. The MacBook achieves only 87-88% accuracy versus Colab's ~98%, with significant slowdowns in versions above 2.12.
  • issues/92564

• Dependency and Installation Challenges: Users face difficulties with TensorFlow installation and dependency management, including specifying newer protobuf versions on Centos 7 and seeking Arch Linux support due to its rolling release model. These challenges highlight the need for clearer installation guidance and broader platform support.
  • issues/92550, issues/92610

• MLIR Conversion and Transformation Errors: Errors occur during MLIR conversion and transformation processes, such as unsupported operations in Voxelmorph model conversion and crashes during "tf-opt" command execution on Ubuntu 20.04. These issues suggest potential bugs or incompatibilities in TensorFlow's MLIR handling.
  • issues/92659, issues/92660

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: 7

Summarized Issues:

• Performance Issues in TensorFlow 2.13.1: Users have reported performance-related problems in TensorFlow version 2.13.1, specifically involving unexpected casting of data types from int32 to float64 during operations. This issue affects the execution of models created using TensorFlow's graph utilities, despite the absence of explicit casting in the source code.
  • issues/90897

• Model Conversion Errors: Several users have encountered errors when converting TensorFlow models to other formats, such as TensorFlow.js and TFLite. These issues include unsupported operations like Erfc and incompatibilities with resource types, suggesting a need for updated conversion tools or model reimplementation.
  • issues/90899, issues/91313

• Build and Install Problems: Users have faced build and installation issues with TensorFlow, such as linking errors on OpenSUSE 15.6 and ImportErrors on Windows 11. These problems are often due to version incompatibilities with compilers, Python, or CUDA/cuDNN, requiring adjustments to the build environment or dependencies.
  • issues/91314, issues/92650

• Attribute and Import Errors: An AttributeError in TensorFlow 2.18 has been reported, where the internal API lacks certain attributes, and an ImportError related to the TensorFlow runtime on Windows 11. Solutions include using public APIs and ensuring compatibility with system dependencies.
  • issues/92116, issues/92650

• Documentation and Badge Link Issues: The TensorFlow project's README file contains broken badge links, leading to 404 error pages. This suggests that the links need to be updated or removed if the target pages no longer exist.
  • issues/92209

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: 5

Key Open Pull Requests

1. feat: add datatype support for tfl.exp and tfl.transpose: This pull request introduces support for the bf16 and f16 data types to the TensorFlow Lite (TFLite) exp and transpose functions, along with corresponding unit tests, to enhance datatype compatibility and functionality within the library.

• URL: pull/92472

• Merged: No

• Associated Commits: 66c76, eb959

2. Fixes #58106: Add shape validation to AvgPool3DGrad input to prevent crash: This pull request addresses issue #58106 by adding shape validation to the AvgPool3DGrad input in the TensorFlow project to prevent potential crashes, as evidenced by updates made in two commits to the avgpooling_op.cc file.

• URL: pull/92521

• Merged: No

• Associated Commits: ae120, 3170a

3. fix(WeakTrackableReference): correct weak reference type handling: This pull request addresses the issue of incorrect handling of weak reference types in the WeakTrackableReference class by implementing a more logical approach to check if the input reference parameter is already a weak reference, converting it if necessary, or using it directly if it is already a weak reference.

• URL: pull/92336

• Merged: No

• Associated Commits: 7b8f5

Other Open Pull Requests

• TensorFlow Allocator Optimization: This pull request addresses an issue in TensorFlow where a BFCAllocator was unnecessarily created when using OneDNN with the TF_ONEDNN_USE_SYSTEM_ALLOCATOR=true setting. The proposed changes aim to prevent this redundant allocator creation, as evidenced by the BFC vlog.
  • pull/92483

• CSV Parsing Enhancement in TensorFlow: This pull request introduces a new feature to TensorFlow's CSV parsing capabilities by creating a CsvDataset.py file. It enhances the parse_csv_line function to support the 'int64' data type alongside the default 'float32', allowing for dynamic handling of columns with different numerical types and improving the library's ability to process datasets with large integer values.
  • pull/92522

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: 4

Key Closed Pull Requests

1. Duplication of ops for which PR already exits: This pull request addresses the addition of missing datatype support for the TensorFlow Lite (TfLite) operations exp, add, and transpose by incorporating support for the bf16 and f16 data types, adding corresponding unit tests, and including non-quantized int8 type support and broadcasting checks for the add operation.

• URL: pull/92469

• Merged: No

• Associated Commits: 3fdab, 85903, a0eba, e4e84

2. Fix pointer stability: This pull request addresses the issue of pointer stability in the TensorFlow project by replacing absl::flat_hash_map, which lacks a pointer stability guarantee and can result in dangling pointers upon rehashing, with std::unique_ptr to ensure stable pointers, and it was successfully merged on April 28, 2025.

• URL: pull/92240

• Merged: 2025-04-28T17:44:09Z

• Associated Commits: 3d199, 87029, 3a1ca

3. Fix typos in documentation strings: This pull request addresses and corrects several typographical errors in the documentation strings of the TensorFlow project, as identified by the contributor, and was successfully merged on May 2, 2025.

• URL: pull/92584

• Merged: 2025-05-02T17:04:49Z

• Associated Commits: 4fa6d

Other Closed Pull Requests

• Spam Pull Request Attempts: This pull request appears to be a non-serious or spam attempt to update the README.md file in the TensorFlow project. It is indicated by its random test description and single commit, and it was not merged.
  • pull/92603

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
mihaimaruseac	2	0	0	12
Venkat6871	2	1	0	11
bchetioui	13	0	0	0
ezhulenev	13	0	0	0
gflegar	8	0	0	0
mtrofin	8	0	0	0
RahulSundarMCW	6	2	0	0
sdasgup3	7	0	0	0
lrdxgm	7	0	0	0