Weekly GitHub Report for Tensorflow: September 22, 2025 - September 29, 2025 (12:06:54)

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 separate 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. ValueError: Only instances ofkeras.Layercan be added to a Sequential model. Received: <tensorflow_hub.keras_layer.KerasLayer object at 0x7f28cb785bd0> (of type <class 'tensorflow_hub.keras_layer.KerasLayer'>): This issue reports a ValueError encountered when attempting to add a tensorflow_hub.keras_layer.KerasLayer instance to a tf.keras.Sequential model, where only instances of keras.Layer are accepted. The user provides a minimal code example that reproduces the error and specifies the TensorFlow and TensorFlow Hub versions used, highlighting a compatibility problem between these components.

   • The comments confirm the issue is reproducible on different platforms and Python versions, and a pull request has been created to address and fix the problem.
   • Number of comments this week: 2

2. XLA GPU JIT: Keras LSTM hits "No registered 'CudnnRNNV3' OpKernel for XLA_GPU_JIT" (TF 2.20.0, Keras 3.11.3): This issue reports a bug where wrapping a Keras LSTM model using the CuDNN fast path with XLA GPU JIT compilation in TensorFlow 2.20.0 results in a compile-time error due to the absence of a registered 'CudnnRNNV3' OpKernel for XLA_GPU_JIT. The problem occurs only when using XLA JIT on GPU, while eager execution works fine, and the user provides a minimal reproducible example demonstrating the failure.

   • The comment explains that the error arises because the default CuDNN-accelerated LSTM is unsupported by XLA GPU JIT, and suggests a workaround by enabling recurrent dropout (setting recurrent_dropout > 0) to disable the CuDNN fast path, allowing successful compilation and execution under XLA on GPU.
   • Number of comments this week: 1

3. [RNN] LSTM model crashing during conversion to int8x8 tflite: This issue describes a problem where an LSTM model crashes during conversion to an int8x8 TensorFlow Lite format on macOS using Keras 3, while the float32 conversion works correctly. The user notes that forcing training mode to false or using Keras 2 avoids the crash but results in poor accuracy, highlighting a compatibility or implementation issue with int8 quantization in the newer Keras version.

   • The single comment indicates a willingness to investigate and fix the int8 quantization crash by creating a minimal reproducer and examining the TFLite converter code.
   • Number of comments this week: 1

4. NotFoundError in cats_and_dogs_filtered tutorial with tf.keras.utils.get_file(extract=True) (TF 2.19, Colab): This issue describes a NotFoundError encountered when running the cats_and_dogs_filtered tutorial using tf.keras.utils.get_file with the extract=True parameter in TensorFlow 2.19 on Google Colab. The error occurs because the tutorial uses an outdated method to construct the file path after extraction, leading to the directory not being found.

   • The single comment indicates that the problem has been resolved through a pull request that updates the tutorial documentation to fix the path construction issue.
   • Number of comments this week: 1

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. TF-TRT Warning: Could not find TensorRT: This issue describes a problem where TensorFlow on Ubuntu 22.04 cannot detect TensorRT despite having an NVIDIA RTX 3050 Ti GPU with the appropriate CUDA 12.4 and NVIDIA driver 535 installed. The user reports difficulties with driver compatibility and has tried multiple reinstallations, but continues to receive a warning that TensorRT could not be found, impacting their ability to use TensorFlow with GPU acceleration.
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 using 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 causes a SystemError and OverflowError. Specifically, when such large uint64 values are passed in eager execution mode, the functions fail with an internal error related to type checking, indicating improper handling of large unsigned 64-bit integers.
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 julius library, into TensorFlow's tf.signal module. The requester highlights the current lack of sophisticated audio data augmentation tools within TensorFlow compared to PyTorch and suggests that adding these capabilities would enhance audio model training by enabling native, efficient preprocessing and augmentation workflows.
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 manually round the real and imaginary parts separately. The user provides a reproducible example and error logs showing that the operation fails due to the absence of a suitable device kernel for complex inputs, highlighting a discrepancy between the official documentation and actual functionality.
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 aborts with a fatal check failure error "Check failed: d < dims()" when invoked with certain random tensor inputs. The problem occurs on Linux Ubuntu 20.04.3 LTS using Python 3.11.8, and the user has confirmed the bug reproduces with TensorFlow Nightly builds, providing a minimal code snippet that triggers the crash.

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: 12

Summarized Issues:

• XLA GPU kernel missing for specific ops: Multiple issues report failures in XLA GPU JIT compilation due to missing registered kernels for certain operations like ImageProjectiveTransformV3, CudnnRNNV3, and RandomPoissonV2. These missing kernels cause compile-time errors and prevent successful execution of functions decorated with @tf.function(jit_compile=True) on GPU.
• issues/100865, issues/100872, issues/100874

• RISC-V platform build and runtime issues: There are problems compiling TensorFlow and TensorFlow Lite natively on the riscv64 architecture, including packaging script failures due to missing arguments and ImportErrors caused by missing shared libraries and undefined symbols. These issues hinder the use of TensorFlow and TFLite on riscv64 systems.
• issues/100940, issues/101017

• LSTM model conversion and quantization crashes: Converting LSTM models to TFLite int8 quantized format causes segmentation faults, while float32 conversions succeed. This crash is potentially linked to Keras 3's enforced training=false behavior during conversion.
• issues/101050

• Data loading path errors in tutorials: Using outdated path construction methods in TensorFlow tutorials leads to NotFoundErrors when loading datasets, as the directory paths are incorrectly referenced. This causes failures in running example code such as the cats_and_dogs_filtered tutorial on platforms like Google Colab.
• issues/101115

• Build failures due to missing Bazel dependencies: TensorFlow builds fail on certain Linux distributions like Void Linux because of missing dependency declarations for protobuf headers in Bazel build rules. This results in compilation errors in core TensorFlow components.
• issues/101142

• XLA JIT compilation failures with string tensors: Enabling XLA JIT compilation causes functions using string tensors to fail because key XLA operations like _Arg, _Retval, and SelectV2 do not support the string data type. This limitation prevents compilation of functions that take, return, or manipulate string tensors under @tf.function(jit_compile=True).
• issues/101181, issues/101183, issues/101188

• TensorFlow Hub and Keras Sequential compatibility issue: Adding a tensorflow_hub.keras_layer.KerasLayer to a tf.keras.Sequential model raises a ValueError because only instances of keras.Layer are accepted. This incompatibility causes problems integrating TensorFlow Hub layers with Keras Sequential models.
• issues/101201

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:

• Spam and inappropriate content: Multiple issues involve users admitting to or reporting spam within the TensorFlow repository, which contributes no valuable information or relevant content to the project. These spam issues disrupt the community and clutter the issue tracker with non-constructive posts.
• issues/100817, issues/100889

• Bug in tf.image.extract_patches function: There is a reported bug in TensorFlow version 2.18.1 where the function tf.image.extract_patches raises a TypeError due to an unexpected keyword argument 'ksizes'. This issue highlights a problem with the function's argument handling when used with custom code.
• issues/100902

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.

1. I am a spammer with no shame, I will spam your project
   • Toxicity Score: 0.75 (Rapid escalation, aggressive language)
   • This GitHub conversation involves a single user expressing frustration towards another user identified as a spammer, with a direct and confrontational tone triggered by unsolicited content.

---

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: 3

Key Open Pull Requests

1. Improved readability and tests: This pull request improves code readability and testing by enhancing tests to cover data types beyond int16, organizing castable data types into a proper container, removing unnecessary f-string usage, and ensuring the changes pass pylint and bazel tests locally.

• URL: pull/100869

• Merged: No

• Associated Commits: 0c6ab, 438a7

2. Fix: near-singular matrix condition number check: This pull request aims to fix the near-singular matrix condition number check by improving the estimation method, addressing issue #94657 in the TensorFlow project.

• URL: pull/100819

• Merged: No

• Associated Commits: 31b4d

3. build(aarch64): Update Compute Library to 52.4.0: This pull request updates the Compute Library from version 24.12 to 52.4.0 for the aarch64 build, incorporating numerous bug fixes and performance improvements, adding a new dependency on a fixed version of KleidiAI, and removing or updating existing Compute Library patches accordingly.

• URL: pull/100882

• Merged: No

• Associated Commits: 0cd9b

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: 1

Key Closed Pull Requests

1. Add XLA kernel support for CropAndResize operation: This pull request addresses TensorFlow issue #100521 by adding a complete XLA kernel implementation for the CropAndResize operation, enabling the tf.image.crop_and_resize function to be compatible with @tf.function(jit_compile=True) through support for both bilinear and nearest neighbor interpolation, proper coordinate transformations, multi-box handling, and registration for CPU and GPU XLA devices without breaking existing API compatibility.

• URL: pull/100897

• Merged: No

• Associated Commits: b3aec

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
ILCSFNO	10	3	9	6
mihaimaruseac	0	0	0	19
zvoicu000	5	3	0	3
Ma-gi-cian	5	2	0	3
adi1220	4	1	0	5
khteh	0	0	5	2
apach301	2	2	2	0
kadirb4rut	3	2	0	1
tinywisdom	0	0	6	0
LiSsHhUuAaIi	0	0	6	0