Weekly GitHub Report for Tensorflow: January 17, 2025 - January 24, 2025
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 v2.18.0
1.2 Version Information:
The TensorFlow 2.18.0 release, created on October 21, 2024, introduces several key updates, including the addition of a fourth parameter to the TfLiteOperatorCreate function for a cleaner API, the disabling of TensorRT support in CUDA builds, and the introduction of Hermetic CUDA for more reproducible builds. Notably, TensorFlow now supports NumPy 2.0 by default, with changes in type promotion rules, and continues to support NumPy 1.26 until 2025. Additionally, the release includes enhancements to tf.data and tf.lite, such as new arguments for map functions and support for TensorType_INT4 and TensorType_INT16 in various operations.
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.
-
Cannot use GpuDelegate - java.lang.IllegalArgumentException: Internal error: Cannot create interpreter: This issue involves a user encountering a
java.lang.IllegalArgumentExceptionerror when attempting to use the GpuDelegate with TensorFlow Lite on an Android device, specifically a Samsung S23. The user has provided a repository to replicate the issue and is seeking assistance in resolving the error, which appears to be related to the configuration or compatibility of the GpuDelegate.- The comments discuss potential configuration issues with the GpuDelegate, suggesting testing with CPU-only execution, which works fine. Suggestions include configuring GpuDelegate options for better control and checking compatibility documentation. The user clarifies the use of GpuDelegate for mobile devices and expresses doubt about the helper's ability to resolve the issue, hoping for input from a more experienced contributor.
- Number of comments this week: 10
-
TFnode on TensorflowonSpark 2.2.5: This issue involves a user experiencing an error with TFNode while running code on Microsoft Fabric using TensorFlow version 2.12 and TensorFlowOnSpark version 2.2.5. The user aims to print a "hello world" message to verify if the code runs on different clusters but encounters compatibility issues due to the mismatch between TensorFlow and TensorFlowOnSpark versions.
- The user initially seeks help for a mind-boggling issue, and a responder requests a minimal code snippet for debugging. Another commenter identifies the problem as a compatibility issue, suggesting using TensorFlow 1.15, but the user cannot downgrade due to platform constraints.
- Number of comments this week: 5
-
It doesn't support on python3.13: This issue is about the inability to install TensorFlow version 2.17 on Python 3.13 due to a lack of support, as indicated by the error message stating no matching distribution was found for TensorFlow. The problem arises because TensorFlow's release cycle does not align with Python's, leading to a delay in support for new Python versions, which has been a recurring issue since Python 3.8.
- The comments discuss the recurring issue of TensorFlow's delayed support for new Python versions, with users expressing frustration over the lack of compatibility with Python 3.13, especially since it is the default version in some major distributions like Fedora 41. Some users suggest downgrading to an older Python version as a temporary solution, while others argue for a more proactive approach from the TensorFlow team to align their release cycle with Python's. The discussion also touches on the complexity of TensorFlow's build system and dependencies, which contribute to the delay in supporting new Python versions.
- Number of comments this week: 4
-
Unable to install TensorFlow: No matching distribution found for TensorFlow!: This issue is about a user who is unable to install TensorFlow due to a "No matching distribution found" error when using Python version 3.13 on Windows 10. The user is attempting to install TensorFlow version 2.8 but encounters compatibility issues with the Python version they are using.
- The comments clarify that TensorFlow 2.8 does not support Python 3.13, and the user is advised to use a compatible Python version, such as 3.8 to 3.11. The user also mentions trying older Python versions like 2.7 and 2.6, which are incompatible with TensorFlow 2.8.
- Number of comments this week: 4
-
Tensorflow 2.14.0 installation/run on C++ in visual studio code: This issue involves a user attempting to set up TensorFlow 2.14.0 for C++ development in Visual Studio Code on macOS 14.4, but encountering errors related to the build process using Bazel. The user is specifically facing difficulties with the target
//tensorflow/tools/pip_package:pip_package, which is not declared, and seeks guidance on how to properly install and configure TensorFlow for C++ usage.- The comments provide guidance on setting up TensorFlow for C++ development, clarifying that the mentioned target is for Python pip packages. Instructions are given for installing dependencies and building the C++ library, but the user encounters a build failure with a specific error message, indicating further troubleshooting is needed.
- Number of comments this week: 4
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.
As of our latest update, there are no stale issues for the project this week.
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: 24
Summarized Issues:
- Compatibility Issues with TensorFlow Versions: Users are experiencing compatibility problems when using TensorFlow with other software or hardware configurations. For instance, TensorFlowOnSpark 2.2.5 is not compatible with TensorFlow 2.12, leading to errors, and downgrading TensorFlow is not an option. Similarly, installing TensorFlow 2.8 on Windows 10 with Python 3.13 fails due to unsupported Python versions, suggesting the need for using a supported Python version.
- Crashes and Errors in TensorFlow Operations: Several TensorFlow operations are causing crashes due to bugs or invalid inputs. Operations like
RaggedTensorToTensor,tf.raw_ops.RaggedGather, andtf.nn.conv3d_transposeare crashing due to segmentation faults, data type mismatches, and invalid argument errors, respectively. These issues are often reproduced with specific inputs and have been reported across different TensorFlow versions.
- Bugs in TensorFlow Pooling and Convolution Functions: TensorFlow's pooling and convolution functions are encountering bugs that lead to crashes. Functions like
tensorflow.nn.max_pool1d,tensorflow.nn.max_pool3d, andtensorflow.compat.v1.nn.depthwise_conv2d_nativeare crashing due to invalid argument errors and issues with input parameters, such as excessively largeksizevalues or invaliddilationsattributes.
- Installation and Build Issues: Users are facing challenges with installing and building TensorFlow on various systems. Problems include failures in installing TensorFlow using Poetry on macOS, building
libtensorflowlite.sofor Android projects, and setting up TensorFlow for C++ development in Visual Studio Code on macOS, often due to missing dependencies or incorrect configurations.
- Feature Requests and Documentation Improvements: There are requests for new features and improvements in TensorFlow's documentation. Users have requested a logging mechanism for GPU memory allocation to better diagnose out-of-memory errors and suggested documentation updates to address installation issues related to CUDA libraries on Linux systems.
- Performance and Optimization Concerns: Users are reporting performance issues and seeking optimization guidance. A TensorFlow tutorial fails due to CUDA device detection errors, and a custom model takes an unusually long time to prepare for training, indicating potential inefficiencies in TensorFlow's handling of specific operations or configurations.
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: 21
Summarized Issues:
- Segmentation Faults in TensorFlow Operations: Segmentation faults have been reported in TensorFlow operations, particularly in
tf.raw_ops.TensorArrayV3on Linux Ubuntu 20.04 when thesizeparameter exceeds system memory limits. This issue causes crashes across different environments with varying memory capacities.
- Discrepancies in Activation Function Outputs: There are discrepancies in output values when using the exponential activation function in TensorFlow, with significant differences between CPU and GPU computations. This issue has been noted to cause large gradient differences during model training, which has been resolved in later versions.
- Crashes and Compatibility Issues in TensorFlow Lite: TensorFlow Lite has encountered crashes and compatibility issues, such as a crash during GRU layer quantization on Apple M1 Pro and a fatal signal error in YOLO11n model on Android. These issues suggest potential bugs or compatibility problems in the TFLite Android library and Keras 3.0.
- Feature Requests and Compatibility Concerns: Feature requests and compatibility concerns have been raised, such as adding int8 support to Unsorted_Segment_X operators for better mobile deployment of GNNs. Additionally, there are requests to publish TensorFlow Lite version 2.18.0 on Maven Central and CocoaPods before transitioning to LiteRT.
- Compilation and Installation Errors: Users have reported various compilation and installation errors, including issues with cross-compiling TensorFlow Lite for RISC systems and errors during TensorFlow C++ interface compilation on Linux. These errors often involve dependency resolution and compatibility with specific Python versions.
- Bugs in TensorFlow and TFLite Kernels: Bugs have been identified in TensorFlow and TFLite kernels, such as inconsistencies in Non-Max Suppression outputs and failures in compiling
tf.keras.layers.Conv2Dwith XLA. These issues lead to slow computation, out-of-memory errors, and requests for fixes to ensure consistent outputs.
- Dependency and Compatibility Issues in Development Environments: Development environments face dependency and compatibility issues, such as Gradle sync failures in the Digit Classifier Codelab and compilation errors on Windows due to incorrect file path handling. These issues often require downgrading or modifying configurations for resolution.
- Documentation and Support Gaps: There are gaps in documentation and support, such as the absence of a compatibility table for TensorFlow 2.18 with CUDA and cuDNN in Spanish documentation. Users have also reported difficulties in understanding installation processes and resolving ImportErrors due to DLL initialization issues.
- Miscellaneous Issues and Requests: Miscellaneous issues include spam removal from the TensorFlow GitHub project and requests for better error reporting in TensorFlow model loading. These issues highlight the need for improved community management and clearer error messages in development tools.
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. All other pull requests are grouped based on similar characteristics for easier analysis.
Pull Requests Opened This Week: 5
Key Open Pull Requests
1. [mlir][tosa] Update Tensorflow to match TOSA v1.0 specification: This pull request aims to update TensorFlow's TOSA (Tensor Operator Set Architecture) implementation to align with the TOSA v1.0 specification by incorporating changes from upstream LLVM patches, specifically addressing updates to convolution operators' accumulator types, TileOp, PadOp, and ensuring equalized ranks for certain operations.
- URL: pull/85608
- Merged: No
2. numpy copy fix: This pull request addresses an issue in the TensorFlow project by modifying the use of the astype() function to prevent unnecessary data copying by propagating copy=None instead of the default copy=True, thereby allowing the function to only perform a copy when explicitly specified.
- URL: pull/85408
- Merged: No
- Associated Commits: 35115
3. 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 with SHA 4dafefa8ac5304c113ad6dc47d5424363002615e, and is currently open for review.
- URL: pull/85639
- Merged: No
- Associated Commits: 4dafe
Other Open Pull Requests
- QNN Types Wrappers in TensorFlow: The pull requests focus on introducing basic wrappers for QNN types in the TensorFlow project. They manage dynamic resources throughout the lifecycle of instances and handle various aspects such as scalar parameters, tensor dimensions, and input-output operations.
- Compiler Replacement with Qualcomm Implementations: These pull requests involve replacing the compiler component with Qualcomm implementations in the TensorFlow project. They include commits from related pull requests, provide instructions for testing the changes, and note that certain models are disabled due to unavailability and one test fails due to a bug in the simple_slice_op.mlir file.
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. All other pull requests are grouped based on similar characteristics for easier analysis.
Pull Requests Closed This Week: 12
Key Closed Pull Requests
1. [RFC] rocprof insights for rocprof data: This pull request proposes the development of a Python package named "rocprof insights" to enhance the analysis and visualization of AMD hardware performance data collected by ROCm/HIP profiling tools, offering functionalities such as data loading, statistical analysis, and visualization through various plots, with potential features like overlaying latency on computational graphs and tracing input/output values for accuracy checks.
- URL: pull/85559
- Merged: No
- Associated Commits: 4ac7a, c2332, 7f4f9, bff30, 1fedf, a1cfa, 3644d, e0aef, be1cc, 07b32, dcac5, f5c6f, ec39c, eeccd, 27d0e, cfe44, f8021, 1ce2d, 32069, 9c01e, 56f62, 8d06b, 84d52, 6af30, cf09f, 4da4f, 47368, cc4a2, b3fcb, b1a42, a3440, 82c8c, c3a1f, 50990, bfe7f, a8220, bca3b, 70706, 1cfeb, 9c905, 50a78, 16101, aae89, ab856, 74698, 60a2d, 18778, 60fc5, fa3a0, abaaa, 184dc, b0af0, c4679, 6d4a5, dc439, f8a8e, 285cc, 096d7, e9be6, 5d5d0, 6a78b, 075b5, 60b8a, 7b52e, 8b7fc, 150bb, f1d1a, 7219d, 426c0, d88d5, 1c607, 0d724, 90403, b7c50, b00c7, 78a7e, 5c652, b488d, ba815, 2dba9, 34d80, ea53c, 40e3e, f65e5, 5a71c, 83713, 70b4d, abb24, 642ae, b6821, 15640, c4958, b6b66, 6d19b, 4c853, d1038, 54e1f, d8198, d860d, c0b36, 18d5c, 7946a, 24c78, 0c992, 2bab5, 89810, 4cc12, 6aa74, 63752, 57188, dc9fb, a91e4, 75ee5, 1e3c6, d40d9, 31d9c, ec39e, 94d31, f42a7, 94f61, 71a8f, d2e43, 1e3ef, ef2d4, 5bad4, 8f1cb, 37d76, e8844, 4556d, cf094, a2036, 33542, 4717f, b9bf3, 3ed87, cd4b0, f40da, a26fb, 038a6, 52994, 79309, 84c13, 4396a, 3c61e, 06702, 6e9f2, da5df, ff760, de1e4, b4e5e, ce1b2, 874c4, 451d1, 8fc12, c1526, 5dbad, 5c1aa, f7991, d377c, 7c927, f5f98, b5400, 1537a, cece9, 7ae16, e624e, 09335, 2211b, 2a929, 6c915, 07cab, 4d444, 575fd, 67b5d, a2bad, 468cf, 7ccb1, 58fc0, db5ae, 9e21f, da24c, 3515d, 8cf48, 09b52, ccced, a382c, 532df, b854f, 58d80, f4770, 3c370, e5dd7, 9f051, 1afb6, f854d, 554f8, 68c43, 2baac, 33a41, e554c, 1c5d7, 4d453, d7f29, 4ca78, 3ac1e, ed3e4, 5788e, 75503, 04c3e, e2acb, 188d1, 1d388, 87930, c1375, 66ca7, b0d33, 713b6, 09053, b6971, d9163, 76f24, 252a9, 7ec4c, 2464a, 8aedd, 071ee, 3a90e, 67d8f, ac4cb, 02686, 25619, e60b7, 70057, fd54d, b74d5, 72a9f, d71ec, f22b6, 8df41, 22772, b1e81, 34e1c, b45da, f8452, 03daf, 2e1af, 09a2b, 6afd0, 88c6b, fff74
2. Branch for Pack and DUS op: This pull request involves multiple updates to the TensorFlow project, including the addition of LiteRt Qualcomm wrappers, modifications to the TensorWrapper to return data sizes in bytes, implementation of a TensorPool for managing TensorWrappers, support for cloning static tensors with different data types, the introduction of op builders, and support for Pack and DUS operations, although it was ultimately not merged.
- URL: pull/85635
- Merged: No
- Associated Commits: 509ec, 7564f, 5a3f2, 451fb, f82a5, c01f6, 94763, 4400f, 21bff, cbc17, 10fcc, a90ee
3. Extend tensorflow restore kernel op to support type casting upon reading.: This pull request aims to enhance the TensorFlow restore kernel operation by adding support for type casting during the reading process, as indicated by the series of commits that include extending the TFRecord dataset, implementing custom protocol buffers, conducting unit tests and optimizations, fixing issues, and specifically extending the restore function to support tensor casting.
- URL: pull/85263
- Merged: No
Other Closed Pull Requests
- Layer Utilities Optimization: This topic focuses on optimizing the layer utilities module by enhancing runtime performance, memory efficiency, and code readability. The improvements include introducing constants for validation, combining validation and conversion steps, improving error handling, adding lazy evaluation, and simplifying input checks.
- Typographical and Link Corrections: This topic addresses the correction of typographical errors and broken links in the TensorFlow project. The corrections were made across multiple files, including
full_type_util.handconfig.proto, and have been successfully merged into the main codebase.
- TensorFlow Installation and Compatibility Fixes: This topic covers updates and fixes related to TensorFlow installation and compatibility issues. The pull requests address installation issues on C++ in Visual Studio Code and ensure compatibility between TensorFlow v2's
ConvertGpuXSpaceToStepStatsfunction and v1'sRunMetadata.
- Matmul Function Correction: This topic involves correcting the misuse of the adjoint parameter for transpose in the vectorized
matmulfunction. The correction ensures accurate results when used withinvectorized_mapand with complex data types, which is crucial for quantum computing simulations.
- Profiler Trace Viewer Display Issue: This topic addresses a display issue with thread IDs in the profiler trace viewer for the TensorFlow project. The pull request aimed to resolve the problem, but it was ultimately not merged.
- Support for Quint8 Type: This topic adds support for the quint8 type to the uniform_quantize and uniform_dequantize operations. The updates include changes to the TF2XLA bridge for proper conversion and involve moving certain operation definitions, with contributions from multiple developers.
- Enhancements to Tosa Conv Ops: This topic focuses on enhancing the Tosa Conv Ops by adding an
acc_typeand adjusting the TileOp to accommodate multiples as input. Despite the proposed enhancements, the pull request was not merged.
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 |
|---|---|---|---|---|
| mihaimaruseac | 6 | 0 | 0 | 78 |
| Venkat6871 | 2 | 2 | 0 | 36 |
| gaikwadrahul8 | 2 | 2 | 0 | 29 |
| tilakrayal | 3 | 2 | 0 | 21 |
| codinglover222 | 7 | 3 | 2 | 4 |
| weilhuan-quic | 11 | 3 | 0 | 0 |
| arzoo0511 | 0 | 0 | 0 | 14 |
| LongZE666 | 0 | 0 | 12 | 1 |
| dnmaster1 | 0 | 0 | 2 | 10 |
| cj401-amd | 9 | 1 | 0 | 0 |