Weekly GitHub Report for Tensorflow: April 21, 2025 - April 28, 2025 (12:03:17)
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.
-
The pip instructions to download and install tensorflow for macOS don't work: This issue highlights a problem with the pip installation instructions for TensorFlow on macOS, where the user is unable to find a compatible version of TensorFlow using the provided commands. The user reports that despite following the official guide, the installation fails with errors indicating no matching distribution is found for both TensorFlow and tf-nightly.
- The comments discuss troubleshooting steps, including verifying the Python and pip versions, and ensuring compatibility with Apple Silicon. There is confusion about the supported Python versions, with discrepancies between the official website and the comments. The user expresses frustration over the lack of visible pip commands in the provided screenshots, and further clarification is given with additional screenshots and advice to follow the official documentation.
- Number of comments this week: 5
-
Tensor Flow not working as expected: This issue reports a bug with TensorFlow version 2.17.0 on an elementary OS 8.0 system, where the user is unable to execute a script using the GPU despite following the ROCm installation guide and setting the necessary environment variables. The user experiences repeated errors related to JIT compilation failure and is seeking guidance to resolve the problem, as the suggested solutions from similar issues have not worked.
- The comments discuss attempts to reproduce the issue, with one user successfully running the code on Colab without issues, suggesting a potential system-specific problem. The original poster clarifies their installation process and expresses difficulty in isolating the issue, while another comment notes that the code works on the CPU with TensorFlow 2.19.0 but not on the GPU, indicating a possible problem with the GPU setup or drivers.
- Number of comments this week: 3
-
Impossible to free GPU memory used by rank 0 tensors: This issue describes a bug in TensorFlow 2.15.1 where rank 0 tensors, once created, cannot be freed from GPU memory, even after using methods like
tf.keras.backend.clear_session()andgc.collect(). The problem persists across different scenarios where rank 0 tensors are created, such as when usingtf.zerosor setting learning rates in optimizers, and is observed on Ubuntu 22.04.5 LTS with CUDA 12.4.- The comments confirm the issue, noting that rank 0 tensors persist in GPU memory despite attempts to clear them, and suggest that the problem might be related to the cudaMallocAsync allocator caching small allocations. A user proposes allowing an opt-out of caching for small tensors or providing an API to release them, while another comment highlights that rank 1 tensors with a single element do not exhibit this issue.
- Number of comments this week: 2
-
module 'tensorflow.python.distribute.input_lib' has no attribute 'DistributedDatasetInterface': This issue involves an AttributeError encountered when using TensorFlow 2.18, specifically related to the absence of the 'DistributedDatasetInterface' attribute in the 'tensorflow.python.distribute.input_lib' module. The error arises when attempting to build neural networks using internal TensorFlow APIs, which have changed in the latest version, causing compatibility issues in PyCharm 2024.
- The comment suggests that the error is due to using internal TensorFlow APIs, which are unstable across versions, and recommends switching to the public Keras API to resolve the issue. It also requests clarification from maintainers on changes to the 'DistributedDatasetInterface' in TensorFlow 2.18 and suggests flagging such breaking changes with deprecation warnings.
- Number of comments this week: 1
-
Broken badge links in README: TF Official Continuous & Nightly return 404: This issue highlights a documentation bug in the TensorFlow project where the badge links for "TF Official Continuous" and "TF Official Nightly" in the README file are broken, leading to a 404 error page. The suggested fix is to update these links or remove them if the target pages no longer exist.
- The comment mentions that the build infrastructure has changed, which is likely the reason the badges were not updated, and tags two individuals who might be responsible for addressing the issue.
- Number of comments this week: 1
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, where the TensorRT library cannot be found despite multiple installation attempts within an Anaconda environment. The user, a graduate student, is facing challenges with the NVIDIA driver compatibility for their RTX 3050 TI GPU, which is compounded by the automatic installation of an unsupported driver version, leading to significant time spent on debugging rather than on their machine learning coursework.
SystemErrorintf.ensure_shapeandtf.compat.v1.ensure_shapewhendtypeofshapeistf.uint64and its value is too large.: This issue involves 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 is reproducible in TensorFlow version 2.15 on Linux Ubuntu 20.04, and it occurs when the APIs are called in eager mode with largeuint64values, such asshape = tf.constant([18446743219011059112, 1], dtype=tf.uint64).- 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 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. - [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 results in an error. The user reports that they must manually apply theRoundoperation to the real and imaginary parts of the tensor separately to achieve the expected behavior, indicating a discrepancy between the documentation and the actual functionality. - 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 "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's handling of tensor dimensions during the unbatching process.
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: 14
Summarized Issues:
- TensorFlow GPU and ROCm Issues: Users are experiencing difficulties with TensorFlow on systems using ROCm, where environment variables fail to resolve GPU-related errors. Despite following installation guides and trying solutions from similar issues, problems like JIT compilation errors and HSACO generation failures persist.
- TensorFlow Functionality and Bug Reports: Several issues highlight bugs and unexpected behaviors in TensorFlow's functionality, such as incorrect comparison results in
WeightedDeltaCompare, inconsistent model outputs, and persistent rank-0 tensors in GPU memory. These issues can lead to crashes, unexpected results, and memory management challenges.
- TensorFlow Installation and Compatibility Problems: Users report issues with TensorFlow installation on macOS and compatibility with new nVidia GPU drivers on Windows. These problems include confusion over Python version requirements and crashes in WSL environments due to driver updates.
- TensorFlow Lite and Model Conversion Challenges: Converting models to TensorFlow Lite and using GPU delegation present challenges, such as non-broadcastable shapes and dynamic shape handling issues. These problems affect inference and delegation processes, particularly with complex models like Vision Transformers.
- TensorFlow API and Documentation Issues: Problems with TensorFlow's API and documentation include missing attributes in internal APIs and broken links in codelabs and README files. These issues hinder user experience and require updates to maintain functionality and accessibility.
- TensorFlow Feature Requests: Users request new features such as a GPU kernel for
tf.linalg.eigoperations and the ability to mute logging in TensorFlow Lite binaries. These requests aim to improve performance and usability for specific use cases.
- TensorFlow Memory Management and Optimization: A bug in TensorFlow 2.15.1 prevents freeing GPU memory due to the default setting of
skip_gradients_aggregation. Adjusting this setting can resolve memory usage issues during gradient descent.
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:
- Installation and Compatibility Issues: Users are experiencing installation failures and compatibility problems with TensorFlow on Windows systems. These issues include errors related to missing DLLs and version mismatches between NumPy versions, which prevent successful execution of custom scripts.
- Documentation Errors: Incorrect documentation in the TensorFlow project misleads users, as highlighted in discussions and related links. This emphasizes the importance of accurate documentation, especially for newcomers to avoid confusion and errors.
- Compilation and Code Bugs: Bugs in TensorFlow include a compilation error due to a missing parenthesis and an AttributeError related to distributed strategies. These issues hinder the development process and require attention to detail in code maintenance.
- TensorFlow Lite Data Type Modifications: Recent changes in TensorFlow Lite's xnnpack delegate altered supported data types for certain functions. This modification affects the tanh and logistic functions, potentially impacting model performance and compatibility.
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: 1
Key Open Pull Requests
1. 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, as detailed in the commits and discussion at https://github.com/tensorflow/tensorflow/pull/92240.
- URL: pull/92240
- Merged: No
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: 8
Key Closed Pull Requests
1. Enable tosa support under conditional compilation: This pull request introduces a clone of tf-opt under tosa, named tf-tosa-opt, to enable TOSA passes through conditional compilation, partially reverting a previous change (#83174) and addressing breaking changes from the MLIR side for better decoupling.
- URL: pull/92055
- Merged: 2025-04-26T19:15:31Z
2. Patch to rules_python to fix python 3.13 free-threading usage: This pull request addresses a temporary patch to the rules_python in order to correctly utilize Python 3.13's free-threading capabilities and ensure proper installation of dependencies built with free-threading, as a workaround until rules_python is upgraded to version 1.X, with references to related issues and pull requests in the JAX and XLA projects.
- URL: pull/91335
- Merged: No
- Associated Commits: fd975
3. Add use_default_shell_env = True to build_pip_package_py rule: This pull request addresses an issue in the TensorFlow project by adding the use_default_shell_env = True parameter to the build_pip_package_py rule, which resolves failures in locating required binaries when they are situated in non-default locations, similar to a previous pull request, and was successfully merged on April 21, 2025.
- URL: pull/91501
- Merged: 2025-04-21T17:18:46Z
- Associated Commits: e4ba5
Other Closed Pull Requests
- TensorFlow Script Fixes: This topic includes a pull request that addresses an issue in TensorFlow's
visualize.pyscript where it was ignoring files with the.tf_litesuffix. The fix ensures these common model files are recognized, although it was not merged into the main codebase.
- Tosa Operator Renaming and Legalization: This topic covers pull requests related to the Tosa operator in TensorFlow. One pull request involves renaming the Tosa operator from "int_div" to "intdiv" and updating the lit tests, while another adds legalization for the
tfl.bitwise_xoroperation in the MLIR TOSA dialect, including tests for supported data types.
- Security Enhancements: This topic includes a pull request that updates URLs to use HTTPS instead of HTTP, enhancing security and providing a safer browsing experience. This change was successfully merged on April 23, 2025.
- XNNPACK Delegate Function Support: This topic involves a pull request that addresses the issue of incorrect supported types for the tanh and logistic functions in the XNNPACK delegate within TensorFlow. The issue was detailed in issue #91896 and the fix was successfully merged on April 23, 2025.
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 | 2 | 0 | 0 | 13 |
| mihaimaruseac | 2 | 0 | 0 | 12 |
| bchetioui | 13 | 0 | 0 | 0 |
| ezhulenev | 13 | 0 | 0 | 0 |
| mtrofin | 8 | 1 | 0 | 2 |
| plopresti | 3 | 1 | 3 | 1 |
| gflegar | 8 | 0 | 0 | 0 |
| sdasgup3 | 7 | 0 | 0 | 0 |
| lrdxgm | 7 | 0 | 0 | 0 |