Weekly GitHub Report for Tensorflow: August 11, 2025 - August 18, 2025 (12:07:04)

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. The tag version of xnnpack is ambiguous: This issue concerns build errors caused by ambiguous or incorrect tag versions of the XNNPACK library used in TensorFlow Lite, where certain types and struct members were missing or incompatible, leading to compilation failures. The user reported errors with multiple XNNPACK commit versions, but a later comment confirmed that updating to the latest synchronized branch of XNNPACK resolved these build issues.

   • The user initially encountered build errors due to mismatches between XNNPACK and TensorFlow Lite code, including missing types and incompatible function signatures. A follow-up comment clarified that using the latest compatible XNNPACK branch fixed these problems, resulting in successful compilation and execution, and suggested closing the issue if resolved.
   • Number of comments this week: 2

2. UserWarning: Protobuf gencode version 5.28.3 is exactly one major version older than the runtime version 6.31.1 at tensorflow/core/...: This issue reports a warning caused by a version mismatch between the protobuf gencode version embedded in TensorFlow 2.20.0 (5.28.3) and the installed protobuf runtime version (6.31.1) on Python 3.13, which triggers a UserWarning about potential future incompatibility. The user seeks support to resolve this warning, which currently does not affect functionality but may cause problems in future protobuf runtime releases.

   • The comments confirm the warning arises because TensorFlow’s protobuf gencode is generated with an older protoc version than the installed runtime, explain the cause and impact, and suggest workarounds such as pinning protobuf to version 5 or ignoring the warning; a request is also made for a minor TensorFlow release to address this by updating the protobuf gencode or adjusting dependencies.
   • Number of comments this week: 2

3. Why does it feel like everything in this space died about 5 years ago? (There's no discussion board): This issue expresses concern over the apparent stagnation and lack of recent development in the Coral NPU and Rockchip hardware space, noting the absence of hardware refreshes and major model updates in several years. The user highlights their own efforts to run a modern YOLO variant on TensorFlow Lite and questions the current state and production use of these technologies, suggesting that the community feels fractured and inactive compared to past enthusiasm.

   • The single comment acknowledges some recent progress with the Rockchip NPU driver being merged into Linux, but agrees that TensorFlow and related projects appear neglected, citing outdated releases and build issues, which contributes to the overall sense of frustration and stagnation in the ecosystem.
   • Number of comments this week: 1

4. Using a tf.Variable as index when writing to a tf.TensorArray yields error in graph mode: This issue describes a bug encountered when using a tf.Variable as an index to write to a tf.TensorArray in TensorFlow’s graph mode, which results in an OperatorNotAllowedInGraphError. The user expected the code to work similarly to eager mode when decorated with @tf.function, but the error arises because Python casts of symbolic tensors are disallowed in graph mode, and the TensorArray was created outside the traced function.

   • The comment explains that the error occurs because Python casting of symbolic tensors (like int(tf.Variable)) is not permitted in graph mode, and passing a TensorArray created in eager mode into a traced function triggers the same error. The recommended solution is to create the TensorArray inside the @tf.function, use a 0-D int tensor as the index (e.g., i.read_value()), and keep index updates symbolic with methods like assign_add. This is a common eager vs. graph mode pitfall rather than user error, and suggestions were made to improve documentation and error messages to clarify these constraints.
   • Number of comments this week: 1

5. ArgMax on CPU does not support complex64/complex128 tensors (NotFoundError): This issue reports that the ArgMax operation on CPU does not support complex64 or complex128 tensors, resulting in a NotFoundError when attempting to use tf.argmax on complex tensors. The user highlights that this limitation is inconsistent with other TensorFlow operations that support complex types and with the XLA kernel registration, suggesting that ArgMax should be extended to handle complex tensors on CPU as well.

   • The single comment acknowledges a misclassification and confirms that complex64 and complex128 types should indeed be supported by ArgMax, implicitly agreeing with the issue's premise.
   • 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.

1. TF-TRT Warning: Could not find TensorRT: This issue describes a problem where the user is encountering a warning that TensorRT cannot be found when running TensorFlow on an Ubuntu 22.04 system with an RTX 3050 Ti GPU and CUDA 12.4. Despite using the NVIDIA driver version 535 and attempting multiple reinstallations, the user is unable to resolve the TensorRT detection error, which is impacting their ability to proceed with machine learning development.
2. 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 reports a bug in TensorFlow where calling tf.ensure_shape or tf.compat.v1.ensure_shape with a shape tensor of type tf.uint64 containing very large values close to 2^64 triggers a SystemError and OverflowError. The problem occurs specifically in eager execution mode, causing these APIs to fail when handling such large shape values, which is unexpected behavior.
3. Feature Request: Integrate different Digital Signal Processing into tf.signal: This issue is a feature request proposing the integration of advanced digital signal processing (DSP) functionalities, similar to those found in the PyTorch julius library, into TensorFlow's tf.signal module. The requester highlights the benefits of having native audio data augmentation tools within TensorFlow to improve model robustness and streamline workflows by reducing reliance on external libraries.
4. [DOCS] Missing complex input for Round op: This issue reports a documentation bug in TensorFlow where the Round operation is described as supporting complex tensor inputs, but in practice, attempting to use a complex tensor with this operation results in an error, requiring users to round the real and imaginary parts separately. The user provides a reproducible example and error logs showing that the operation fails due to a missing device kernel for complex inputs, highlighting a discrepancy between the official documentation and the actual behavior in TensorFlow version 2.15.0.
5. tf.raw_ops.Unbatch aborts with "Check failed: d < dims()": This issue reports a bug in TensorFlow version 2.17 where the operation tf.raw_ops.Unbatch causes the program to abort with a "Check failed: d < dims()" error. The problem occurs when running custom code on Linux Ubuntu 20.04.3 LTS with Python 3.11.8, and it has been reproduced using TensorFlow Nightly, indicating a potential flaw in handling tensor dimensions within the Unbatch operation.

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:

• XNNPACK Tag and Build Errors: Multiple issues report build failures and errors caused by incorrect or ambiguous XNNPACK tag versions in TensorFlow Lite, where the specified commits do not exist or point to wrong repositories, leading to missing types and incompatible code that prevent successful compilation. These problems affect both the TensorFlow Lite CMake module and the build_pip_package_with_cmake.sh command, causing significant build disruptions.
• issues/98662, issues/98772, issues/98967

• Coral NPU and Edge AI Ecosystem Stagnation: There is concern over the lack of recent development and active community engagement in the Coral NPU and related hardware/software ecosystems, with no new models, hardware refreshes, or significant updates in the TensorFlow Lite and edge AI space over several years. This stagnation raises questions about the future viability and support for these platforms.
• issues/98783

• TensorFlow Graph Mode Indexing Bug: A bug causes an OperatorNotAllowedInGraphError when using a tf.Variable as an index to write to a tf.TensorArray in graph mode, due to disallowed Python casts of symbolic tensors. The issue highlights the need to create the TensorArray inside a @tf.function and use a scalar tensor index instead of a Python integer cast.
• issues/98814

• Flutter-tflite Plugin Android 15 Compatibility: There is a request for support of 16KB page sizes in the flutter-tflite plugin to comply with upcoming Android 15 requirements mandating all shared libraries be recompiled for 16KB page size compatibility. This change is necessary to ensure continued functionality on new Android versions.
• issues/98933

• TensorFlow Lite Converter Weight Stripping Bug: TensorFlow Lite’s converter strips weights loaded from external .h5 files during custom tf.Module initialization for on-device training, resulting in drastically smaller converted models and runtime READ_VARIABLE errors due to missing variable initializations. This bug severely impacts model accuracy and usability after conversion.
• issues/98964

• Protobuf Version Mismatch Warning: TensorFlow embeds protobuf gencode version 5.28.3, which is one major version older than the installed protobuf runtime 6.31.1, causing UserWarnings on Python 3.13 systems. This version mismatch suggests TensorFlow should update its protobuf gencode to prevent future compatibility issues.
• issues/98980

• Layout Optimization Size Mismatch Error: An INVALID_ARGUMENT error occurs during layout optimization in TensorFlow 2.20.0 due to a size mismatch between values and permutation in the stateless dropout operation within a GAN tutorial. Although the model continues to run, the error raises concerns about its source, impact, and potential fixes.
• issues/98985

• ArgMax Operation Lacks Complex Tensor Support on CPU: The TensorFlow CPU implementation of ArgMax does not support complex64 or complex128 tensors, resulting in a NotFoundError, despite other operations and the XLA kernel supporting these types. There is a call to extend ArgMax to handle complex tensors on CPU for consistency.
• issues/98988

• SparseTensor Support Missing in Math Operations: TensorFlow math operations like tf.asinh do not support SparseTensor inputs and raise a ValueError instead. Users request either direct SparseTensor support or clearer documentation requiring conversion to dense tensors.
• issues/98994

• tf.audio.decode_wav Crash on Invalid Parameter: Calling tf.audio.decode_wav with an invalid desired_channels parameter (e.g., negative value) causes an immediate C++ runtime crash instead of a proper Python exception, leading to an ungraceful process abort. This bug affects error handling robustness.
• issues/98996

• TFLite GPU Delegate Fails on Dynamic-Sized Tensors: The MobileViT model converted to TensorFlow Lite runs successfully on CPU and XNNPACK but fails with the TFLite GPU delegate due to unsupported dynamic-sized tensors. This limitation restricts GPU delegate usage for certain models.
• issues/99047

• INT8 Quantized Model Accuracy Drop: Converting a Keras .h5 model to TensorFlow Lite INT8 format results in severely biased predictions and significant accuracy drops compared to float16 conversion, despite verified data loading and testing. This issue questions the reliability of INT8 quantization for some models.
• issues/99052

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: 3

Summarized Issues:

• TensorFlow runtime errors: Several issues involve runtime errors in TensorFlow, including a NotImplementedError when converting symbolic tensors to numpy arrays during training and an ImportError due to DLL load failure on Windows 11 with Python 3.11.4. These errors highlight problems with TensorFlow's compatibility and functionality in specific environments and use cases.
• [issues/98779, issues/98922]

• User behavior and moderation: One issue involves a user self-identifying as a spammer and referencing unrelated projects and comments, indicating concerns about user conduct and moderation within the TensorFlow GitHub repository. This points to challenges in managing community interactions and maintaining issue quality.
• [issues/98920]

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. Fix logic error in _none_to_default function: This pull request fixes a logic error in the _none_to_default function by ensuring it returns the actual inputs value when inputs is not None, and only returns the default value when inputs is None.

• URL: pull/98644

• Merged: No

• Associated Commits: dad28

2. Improve grammar in regularizer error message: This pull request improves the grammar in the error message related to invalid regularization penalty numbers by changing the phrase "is not a property value" to "is not a valid property value" for enhanced clarity and correctness.

• URL: pull/98645

• Merged: No

• Associated Commits: 9581e

3. Fix grammar in regex_replace docstring: This pull request aims to improve the clarity and readability of the regex_replace function's docstring by fixing a grammatical error, specifically changing the phrase "can be to insert" to "which can be used to insert."

• URL: pull/98646

• Merged: No

• Associated Commits: 31cba

Other Open Pull Requests

• Build and Compatibility Fixes: These pull requests address compatibility issues with different libc implementations and hardware architectures. One fixes build issues by adding the GLIBC macro to conditionally compile functions specific to GLIBC, while the other resolves misaligned access problems in GPU kernels by enforcing stricter memory alignment for improved stability and performance on newer GPUs.
• pull/98766, pull/99046

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: 3

Key Closed Pull Requests

1. r2.19 cherry-pick: ead4f60e7e6 "Reverts 6aad90114b5a78885b3087741992ffc496a79127": This pull request reverts a previous commit identified by SHA 6aad90114b5a78885b3087741992ffc496a79127 in the TensorFlow repository, effectively undoing the changes introduced by that commit.

• URL: pull/98688

• Merged: Yes

• Associated Commits: 8d0d8

2. [oneDNN] Change public TensorFlow CPU build configuration for avx_linux and avx_win: This pull request modifies the public TensorFlow CPU build configuration for both Linux and Windows by adding the Bazel option "--copt='-mf16c'" to enable the f16c instruction set, which improves fp16 performance starting with Ivy Bridge processors.

• URL: pull/98786

• Merged: Yes

• Associated Commits: 08a49

3. Fix build with musl: This pull request addresses build issues with musl by adding the GLIBC macro to conditionally include the 'mallinfo()' and 'mallinfo2()' functions, which are specific to GLIBC and not available in other libc implementations like musl.

• URL: pull/98720

• Merged: No

• 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.

Contributor	Commits	Pull Requests	Issues	Comments
tensorflower-gardener	185	0	0	0
mihaimaruseac	0	0	0	63
ezhulenev	35	0	0	0
djxnqp	0	0	31	0
Aaraviitkgp	10	3	0	7
WillFroom	13	0	0	0
No author found	12	0	0	0
beckerhe	12	0	0	0
wsmoses	12	0	0	0
mkuperst	12	0	0	0