CVE-2021-37651

TensorFlow is an end-to-end open source platform for machine learning. In affected versions the implementation for `tf.raw_ops.FractionalAvgPoolGrad` can be tricked into accessing data outside of bounds of heap allocated buffers. The [implementation](https://github.com/tensorflow/tensorflow/blob/f24faa153ad31a4b51578f8181d3aaab77a1ddeb/tensorflow/core/kernels/fractional_avg_pool_op.cc#L205) does not validate that the input tensor is non-empty. Thus, code constructs an empty `EigenDoubleMatrixMap` and then accesses this buffer with indices that are outside of the empty area. We have patched the issue in GitHub commit 0f931751fb20f565c4e94aa6df58d54a003cdb30. The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, TensorFlow 2.4.3, and TensorFlow 2.3.4, as these are also affected and still in supported range.

• CWE-125
• CWE-787

• https://github.com/tensorflow/tensorflow/commit/0f931751fb20f565c4e94aa6df58d54a003cdb30Patch, Third Party Advisory
• https://github.com/tensorflow/tensorflow/security/advisories/GHSA-hpv4-7p9c-mvfrThird Party Advisory
• https://github.com/tensorflow/tensorflow/commit/0f931751fb20f565c4e94aa6df58d54a003cdb30Patch, Third Party Advisory
• https://github.com/tensorflow/tensorflow/security/advisories/GHSA-hpv4-7p9c-mvfrThird Party Advisory

Published

Aug 12, 2021

Last Modified

Jun 17, 2026

Status

Modified

• How Strix found a critical auth bypass in etcdStrix autonomously discovered a critical authentication bypass in etcd, later designated CVE-2026-33413.
• Autonomous PentestingAI agents that find and validate exploitable vulnerabilities like this one across your applications.
• PR ReviewsPentest every pull request so vulnerable code is caught before it ships to production.
• AI Penetration TestingHow AI-driven penetration testing continuously covers your attack surface.

• CVE-2021-37646TensorFlow is an end-to-end open source platform for machine…5.5
• CVE-2021-37647TensorFlow is an end-to-end open source platform for machine…5.5
• CVE-2021-37648TensorFlow is an end-to-end open source platform for machine…7.8
• CVE-2021-37649TensorFlow is an end-to-end open source platform for machine…5.5
• CVE-2021-3765validator.js is vulnerable to Inefficient Regular Expression…7.5
• CVE-2021-37650TensorFlow is an end-to-end open source platform for machine…7.8
• CVE-2021-37652TensorFlow is an end-to-end open source platform for machine…7.8
• CVE-2021-37653TensorFlow is an end-to-end open source platform for machine…5.5
• CVE-2021-37654TensorFlow is an end-to-end open source platform for machine…7.1
• CVE-2021-37655TensorFlow is an end-to-end open source platform for machine…7.3
• CVE-2021-37656TensorFlow is an end-to-end open source platform for machine…7.8
• CVE-2021-37657TensorFlow is an end-to-end open source platform for machine…7.8