tensorflow
Affected
- introduced
0, fixed2.6.4(ecosystem)
Fixed versions
2.6.4
Fix
- Fixed version
2.6.4pip install --upgrade tensorflow==2.6.4
Advisory
Also known as: BIT-tensorflow-2022-29212, CVE-2022-29212, PYSEC-2026-3141, PYSEC-2026-3299, PYSEC-2026-972
Core dump when loading TFLite models with quantization in TensorFlow
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.
Medium Basis: The source's own band: MODERATE
0, fixed 2.6.4 (ecosystem)2.6.4
2.6.4pip install --upgrade tensorflow==2.6.42.7.0, fixed 2.7.2 (ecosystem)2.7.2
2.7.2pip install --upgrade tensorflow==2.7.22.8.0, fixed 2.8.1 (ecosystem)2.8.1
2.8.1pip install --upgrade tensorflow==2.8.10, fixed 2.6.4 (ecosystem)2.6.4
2.6.4pip install --upgrade tensorflow-cpu==2.6.42.7.0, fixed 2.7.2 (ecosystem)2.7.2
2.7.2pip install --upgrade tensorflow-cpu==2.7.22.8.0, fixed 2.8.1 (ecosystem)2.8.1
2.8.1pip install --upgrade tensorflow-cpu==2.8.10, fixed 2.6.4 (ecosystem)2.6.4
2.6.4pip install --upgrade tensorflow-gpu==2.6.42.7.0, fixed 2.7.2 (ecosystem)2.7.2
2.7.2pip install --upgrade tensorflow-gpu==2.7.22.8.0, fixed 2.8.1 (ecosystem)2.8.1
2.8.1pip install --upgrade tensorflow-gpu==2.8.1Fix commands are the export's own generic commands. Ojuju never runs them. Check them against your setup first.
Check your own computer: guard packages check
Source: GitHub Advisory Database (reviewed advisories) · CC BY 4.0