Random Output Vulnerability in TensorFlow by Google
CVE-2025-55556
6.5MEDIUM
What is CVE-2025-55556?
A vulnerability in TensorFlow v2.18.0 allows the framework to produce random results when compiling the Embedding function. This unexpected behavior can lead to inconsistent application performance and results, potentially compromising the integrity of applications that rely on stable output. Developers are advised to review their code and check for the use of affected versions to mitigate any risks associated with this behavior.