Core dump when loading TFLite models with quantization in TensorFlow

GHSA-8wwm-6264-x792 · BIT-tensorflow-2022-29212 · CVE-2022-29212 · PYSEC-2026-3141 · PYSEC-2026-3299 · PYSEC-2026-972 · 갱신 2026. 7. 14.

생태계	패키지	수정 버전
PyPI	tensorflow	2.6.42.7.22.8.1
PyPI	tensorflow-cpu	2.6.42.7.22.8.1
PyPI	tensorflow-gpu	2.6.42.7.22.8.1

Impact

Certain TFLite models that were created using TFLite model converter would crash when loaded in the TFLite interpreter. The culprit is that during quantization the scale of values could be greater than 1 but code was always assuming sub-unit scaling.

Thus, since code was calling QuantizeMultiplierSmallerThanOneExp, the TFLITE_CHECK_LT assertion would trigger and abort the process.

Patches

We have patched the issue in GitHub commit a989426ee1346693cc015792f11d715f6944f2b8.

The fix will be included in TensorFlow 2.9.0. We will also cherrypick this commit on TensorFlow 2.8.1, TensorFlow 2.7.2, and TensorFlow 2.6.4, as these are also affected and still in supported range.

For more information

Please consult our security guide for more information regarding the security model and how to contact us with issues and questions.

Attribution

This vulnerability has been reported externally via a GitHub issue.

Core dump when loading TFLite models with quantization in TensorFlow

Core dump when loading TFLite models with quantization in TensorFlow

영향 받는 패키지

원문

Impact

Patches

For more information

Attribution

출처