CVE-2025-55559: TensorFlow Denial of Service Vulnerability in Conv2D Layer
Summary
CVE-2025-55559: An issue was discovered TensorFlow v2.18.0. A Denial of Service (DoS) occurs when padding is set to 'valid' in tf.keras.layers.Conv2D. Affects tensorflow versions >= 2.18.0, < 2.18.1. Fixed in 2.18.1. How to fix ↓
Check if this vulnerability affects your application
Upload a manifest or SBOM and we'll show whether you're affected, plus every other risky package in the same scan.
No signup required · Results in about 60 seconds
Background
TensorFlow is an open-source machine learning framework developed by Google, widely used for building and training machine learning models. It provides a comprehensive ecosystem for deep learning, including tools for model deployment and optimization. The framework's Conv2D layer is essential for processing image data, allowing for convolutional operations that are fundamental in computer vision tasks. However, a vulnerability has been identified in version 2.18.0, which can lead to a Denial of Service (DoS) when specific padding configurations are used.
Vulnerability Detail
The vulnerability arises in TensorFlow version 2.18.0, specifically within the tf.keras.layers.Conv2D layer. When the padding is set to 'valid', it can trigger a Denial of Service (DoS) condition. This issue can be exploited by sending specially crafted input data to the Conv2D layer, potentially causing the application to become unresponsive or crash. The severity of this vulnerability is classified as high, indicating that it poses a significant risk to applications utilizing this version of TensorFlow.
How To Fix
Minimum fixed version
2.18.1
This is the first release that contains the fix.
Tip — Not sure which version your app runs? Vulert can match this fix to it. Scan free →
Workaround
If immediate upgrading is not feasible, consider changing the padding parameter in the Conv2D layer from 'valid' to 'same' to mitigate the risk of triggering the DoS condition. Example code: tf.keras.layers.Conv2D(filters=32, kernel_size=(3, 3), padding='same').
Conclusion
The CVE-2025-55559 vulnerability in TensorFlow poses a significant risk of Denial of Service when using the Conv2D layer with 'valid' padding. It is crucial to upgrade to version 2.18.1 or later to mitigate this risk. For those unable to upgrade immediately, adjusting the padding parameter serves as a temporary workaround.
See every vulnerability recorded for tensorflow →
FAQ
CVE-2025-55559 is a vulnerability in TensorFlow that can lead to a Denial of Service when using the Conv2D layer with 'valid' padding.
TensorFlow version 2.18.0 is affected by this vulnerability. Version 2.18.1 and later contain the fix.
You can use Vulert, which monitors open-source vulnerabilities in real-time. Simply upload your manifest file like package-lock.json to check for vulnerabilities.
As a temporary measure, change the padding parameter in the Conv2D layer to 'same' to avoid the DoS issue.
Detailed information can be found in the vulnerability database at https://github.com/bitnami/vulndb/tree/main/data/tensorflow/BIT-tensorflow-2025-55559.json.
References
Related vulnerabilities
Recent vulnerabilities recorded for tensorflow.
