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:

Released on March 5, 2025, TensorFlow version 2.19.0 introduces breaking changes to the tf.lite API, including the deprecation of tf.lite.Interpreter in favor of ai_edge_litert.interpreter and changes to certain C++ constants for improved API compatibility. Key updates also include runtime support for the bfloat16 data type in the tfl.Cast operation and the discontinuation of standalone libtensorflow package publishing, while still allowing unpacking from PyPI.

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. [TYPE:BUG] [COMP:XLA] [TF 2.19] XLA compilation failure for tf.random.shuffle: Compare type mismatch (SIGNED vs TOTALORDER) with int32 operand: This issue reports an XLA compilation failure when using tf.random.shuffle with jit_compile=True due to a comparison type mismatch error involving SIGNED versus TOTALORDER with an int32 operand, while the same code works fine in eager mode. The problem stems from XLA's inability to compile tf.random.shuffle because of its stateful random number generation and unsupported comparison semantics during sorting operations.

   • The comments confirm the issue is reproducible across platforms and TensorFlow versions, explain that the error arises from XLA's internal handling of stateful RNG and sorting operations, suggest a workaround by performing the shuffle outside the compiled function, and discuss the lack of official documentation on unsupported ops and the need for clearer error messages; contributors express interest in fixing the problem at the XLA compiler level and clarify that some stateful random ops are supported while others like shuffle are not.
   • Number of comments this week: 6

2. [TYPE:BUG] [COMP:LITE] TFLite C++ /Python CPU usage error for num_threads > 1: This issue describes a problem where TensorFlow Lite's C++ implementation on a Raspberry Pi 5 does not properly utilize multiple CPU threads when running inference with the XNNPACK delegate, resulting in significantly lower CPU usage and slower performance compared to the Python implementation. The user has confirmed building from the master branch with XNNPACK enabled, but observes that only one thread is heavily used in C++ while all threads are effectively utilized in Python.

   • The comments focus on verifying the build configuration, specifically whether XNNPACK support was enabled during compilation, with the user confirming that it was enabled and that threads are created but not properly utilized, indicating a potential issue with thread management in the C++ API.
   • Number of comments this week: 4

3. [TYPE:BUG] [AWAITING PR MERGE] tf.reverse in eager mode does not validate axis range for scalar tensors (graph mode correctly errors): This issue reports a bug in TensorFlow where the tf.reverse function in eager execution mode does not properly validate the axis range for scalar tensors, allowing invalid axes without raising an error, whereas graph mode correctly raises a ValueError. The user provides a minimal reproducible example demonstrating that eager mode returns the input unchanged when given out-of-range axes, while graph mode throws the expected error, and suggests adding explicit axis validation in the eager execution path to fix this inconsistency.

   • The comments discuss the root cause of the discrepancy, identifying that eager mode bypasses certain validations present in graph mode, and propose a solution involving adding axis range checks in the Python wrapper for tf.reverse. A contributor confirms submitting a pull request to implement this fix and requests feedback on the changes.
   • Number of comments this week: 2

4. [STAT:AWAITING RESPONSE] [TYPE:BUG] [TF 2.19] Footer language dropdown overflows viewport and becomes unselectable across tensorflow.org: This issue describes a bug where the language selector dropdown in the footer of tensorflow.org overflows outside the visible viewport, making it difficult or impossible for users to view or select language options. The problem occurs across multiple devices and screen sizes, including desktop browsers and mobile responsive views, and has been reproduced on TensorFlow version 2.20.

   • A comment suggests trying the language selector dropdown located at the top of the page as an alternative, accompanied by a screenshot for clarification, but no further discussion or resolution has been provided yet.
   • Number of comments this week: 1

5. [TYPE:PERFORMANCE] [TF 2.19] Basic import tensorflow takes 10seconds: This issue reports that importing the TensorFlow library in Python takes an unusually long time—around 10 seconds—on a Windows 11 system with Python 3.9, despite similar hardware and software configurations where the import is faster. The user has provided detailed logs and a verbose output to help diagnose the performance bottleneck during the import process.

   • A single comment was made sharing a profiler output file to assist in analyzing the slow import time, contributing additional diagnostic data to the investigation.
   • 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.

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: 8

Summarized Issues:

• Function behavior inconsistencies: Several issues highlight inconsistent behavior in TensorFlow functions across different modes or platforms. For example, tf.reverse does not validate axis range for scalar tensors in eager mode, and tf.linalg.band_part inconsistently enforces argument validation between eager CPU and TF-XLA modes, leading to unexpected results or silent failures.
• [issues/110038, issues/110316]

• Performance and threading issues: TensorFlow exhibits performance problems on specific platforms and configurations. Importing TensorFlow 2.20.0 on Windows 11 with Python 3.9 is unusually slow, and TensorFlow Lite's C++ implementation on Raspberry Pi5 fails to utilize multiple CPU threads properly, causing slower inference performance compared to Python.
• [issues/110143, issues/110160]

• Compilation and type errors: Using XLA compilation with certain TensorFlow operations can cause failures due to type mismatches. Specifically, tf.random.shuffle with jit_compile=True triggers a type mismatch error between SIGNED and TOTALORDER comparison types on int32 operands, an issue not present in eager mode.
• [issues/110219]

• User interface and usability problems: The language selector dropdown on tensorflow.org's footer overflows outside the viewport on multiple devices, preventing users from properly viewing or selecting language options. This UI bug affects accessibility and user experience across screen sizes.
• [issues/110139]

• Security and dependency concerns: There is a security vulnerability in the protobuf library used by TensorFlow, where the required protobuf version is less than 5 despite fixes being available in protobuf 5.29.6. This raises concerns about the impact of the vulnerability on TensorFlow and requests for updated releases supporting protobuf version 6 or clarification on the issue.
• [issues/110343]

• Build and migration challenges: TensorFlow requires a gradual migration plan to move away from legacy Android crosstool flags and config settings. This migration aims to improve compatibility with Bazel 7+ defaults and reduce maintenance burdens on downstream projects by adopting platform and toolchain resolution.
• [issues/110371]

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: 2

Summarized Issues:

• XLA Compilation Errors with Specific Inputs: There are issues where certain operations or model names that work fine in eager execution mode cause errors during XLA compilation with jit_compile=True. These errors include a ValueError due to invalid root scope name validation and a TypeError caused by a type mismatch when using a tensor as a seed parameter in dropout operations.
• issues/108501, issues/108502

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: 14

Key Open Pull Requests

1. Fix issue #110038: Validate axis range for scalar tensors in tf.reverse: This pull request fixes the eager mode behavior of tf.reverse by validating the axis range for scalar tensors before the scalar check, ensuring that invalid axes raise an error consistently across eager and graph execution modes, and includes comprehensive test coverage for scalar tensor edge cases.

• URL: pull/110061

• Associated Commits: a9ee0, 0c5d4, 0c808, 9a3c8

2. Cleanup/ops todo deprecation removal: This pull request focuses on cleaning up the TensorFlow codebase by removing deprecated TODO comments and dead code in ops.py, replacing deprecated functions with modern runnable doctest examples across multiple modules, eliminating unused parameters and private aliases, and converting non-runnable code blocks in documentation into proper runnable doctests to improve code clarity and maintainability.

• URL: pull/110149

• Associated Commits: d3d24, 0106a, aea17, 4b7d9

3. Add type annotations and comprehensive docstrings to tensorflow/python/types/core.py: This pull request adds detailed type annotations and comprehensive docstrings to the core TensorFlow type system in tensorflow/python/types/core.py, enhancing code clarity, developer experience, and IDE support by addressing existing TODOs and providing practical examples for protocol implementation.

• URL: pull/109974

• Associated Commits: 6acd3, b3e15

Other Open Pull Requests

• Axis validation and error handling in tensor operations: Multiple pull requests address improving axis parameter validation and error reporting in TensorFlow operations. These include adding Python-level axis validation for tf.reverse in eager mode to raise ValueError for invalid axes, fixing confusing errors in tf.squeeze by raising TypeError for tensor axes, and moving axis validations earlier in Reverse kernels to ensure consistent error raising across modes.
• pull/110060, pull/110201, pull/110310

• Documentation improvements and contribution guidelines: Several pull requests focus on enhancing documentation and contributor resources. These include a comprehensive cleanup of 107 markdown files fixing typos, grammar, and formatting, adding runnable doctest examples and fixing broken links, and introducing new configuration files like protegidas.yml and Contributing.yml along with updates to the CODE_OF_CONDUCT.md and README to clarify contribution processes and standards.
• pull/110107, pull/110135, pull/110136, pull/110140, pull/110174

• Code cleanup and static type checking fixes: Some pull requests improve code quality and static analysis accuracy. This includes removing dead code and unused imports in function.py related to the always-enabled C API path, and fixing false type errors reported by Pylance and Pyright by correcting return type annotations and adding type stubs for Tensor and Variable classes.
• pull/110146, pull/110383

• Quantization type correction: One pull request fixes the supported quantization type check for i16 quantization by correcting it from signed int16 to signless int16. This change enables proper handling in the uniform_quantized_stablehlo_to_tfl_pass.cc and updates a related unit test that was previously passing incorrectly.
• pull/110395

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: 2

Key Closed Pull Requests

1. Add tensorflow_text to the repo: This pull request proposes adding the original tensorflow_text library at version 2.20.0 to the repository without any code modifications, except for removing git-related files specific to tensorflow_text and relocating its code files into the core folder.

• URL: pull/110029

• Associated Commits: eb82f, 6194d

• Associated Commits: eb82f, 6194d

2. Add type annotations and comprehensive docstrings to tensorflow/python/types/core.py: This pull request aims to enhance the tensorflow/python/types/core.py module by adding detailed type annotations and comprehensive docstrings to improve code clarity, developer experience, and static type checking, while also addressing existing TODOs and providing practical examples for better understanding of the TensorFlow core type system.

• URL: pull/109973

• Associated Commits:

• Associated Commits:

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.

Access Last Week's Newsletter:

• Link