Integer overflow in `SpaceToBatchND`

ImpactThe implementation of tf.rawops.SpaceToBatchND in all backends such as XLA and handwritten kernels is vulnerable to an integer overflow:pythonimport tensorflow as tfinput = tf.constant-3.5e+35, shape=10,19,22, dtype=tf.float32blockshape = tf.constant-1879048192, shape=2,...

PYSEC-2021-723

TensorFlow is an end-to-end open source platform for machine learning. The implementation of the SpaceToBatchNd TFLite operator is vulnerable to a division by zero error. An attacker can craft a model such that one dimension of the block input is 0. Hence, the corresponding value in blockshape is...

PT-2021-18348 · Google · Tensorflow

Name of the Vulnerable Software and Affected Versions: TensorFlow versions prior to 2.5.0 TensorFlow versions 2.4.2, 2.3.3, 2.2.3, and 2.1.4 are also affected Description: The implementation of the SpaceToBatchNd TFLite operator is vulnerable to a division by zero error. An attacker can craft a...