FPE in `tf.image.generate_bounding_box_proposals`

Impact When running on GPU, tf.image.generateboundingboxproposals receives a scores input that must be of rank 4 but is not checked. python import tensorflow as tf a = tf.constantvalue=1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0 b = tf.constantvalue=1...

GHSA-8FVV-46HW-VPG3 Overflow in `tf.keras.losses.poisson`

Impact tf.keras.losses.poisson receives a ypred and ytrue that are passed through functor::mul in BinaryOp. If the resulting dimensions overflow an int32, TensorFlow will crash due to a size mismatch during broadcast assignment. python import numpy as np import tensorflow as tf truevalue =...

Overflow in `tf.keras.losses.poisson`

Impact tf.keras.losses.poisson receives a ypred and ytrue that are passed through functor::mul in BinaryOp. If the resulting dimensions overflow an int32, TensorFlow will crash due to a size mismatch during broadcast assignment. python import numpy as np import tensorflow as tf truevalue =...

GHSA-54PP-C6PP-7FPX Overflow in `ImageProjectiveTransformV2`

Impact When tf.rawops.ImageProjectiveTransformV2 is given a large output shape, it overflows. python import tensorflow as tf interpolation = "BILINEAR" fillmode = "REFLECT" images = tf.constant0.184634328, shape=2,5,8,3, dtype=tf.float32 transforms = tf.constant0.378575385, shape=2,8,...

Overflow in `ImageProjectiveTransformV2`

Impact When tf.rawops.ImageProjectiveTransformV2 is given a large output shape, it overflows. python import tensorflow as tf interpolation = "BILINEAR" fillmode = "REFLECT" images = tf.constant0.184634328, shape=2,5,8,3, dtype=tf.float32 transforms = tf.constant0.378575385, shape=2,8,...

Overflow in `FusedResizeAndPadConv2D`

Impact When tf.rawops.FusedResizeAndPadConv2D is given a large tensor shape, it overflows. python import tensorflow as tf mode = "REFLECT" strides = 1, 1, 1, 1 padding = "SAME" resizealigncorners = False input = tf.constant147, shape=3,3,1,1, dtype=tf.float16 size =...

Seg fault in `ndarray_tensor_bridge` due to zero and large inputs

Impact If a numpy array is created with a shape such that one element is zero and the others sum to a large number, an error will be raised. E.g. the following raises an error: python np.ones0, 231, 231 An example of a proof of concept: python import numpy as np import tensorflow as tf inputval =...

GHSA-JQ6X-99HJ-Q636 Seg fault in `ndarray_tensor_bridge` due to zero and large inputs

Impact If a numpy array is created with a shape such that one element is zero and the others sum to a large number, an error will be raised. E.g. the following raises an error: python np.ones0, 231, 231 An example of a proof of concept: python import numpy as np import tensorflow as tf inputval =...

GHSA-W58W-79XV-6VCJ Out of bounds segmentation fault due to unequal op inputs in Tensorflow

Impact tf.rawops.DynamicStitch specifies input sizes when it is registered. cpp REGISTEROP"DynamicStitch" .Input"indices: N int32" .Input"data: N T" .Output"merged: T" .Attr"N : int = 1" .Attr"T : type" .SetShapeFnDynamicStitchShapeFunction; When it receives a differing number of inputs, such as...

Out of bounds segmentation fault due to unequal op inputs in Tensorflow

Impact tf.rawops.DynamicStitch specifies input sizes when it is registered. cpp REGISTEROP"DynamicStitch" .Input"indices: N int32" .Input"data: N T" .Output"merged: T" .Attr"N : int = 1" .Attr"T : type" .SetShapeFnDynamicStitchShapeFunction; When it receives a differing number of inputs, such as...

Denial Of Service (DoS)

tensorflow is vulnerable to denial of service. The vulnerability exists in the FusedResizeAndPadConv2D function of nnops.cc due to improper buffer size checking which allows an attacker to cause an application crash by providing malicious input...

PT-2022-26134

Name of the Vulnerable Software and Affected Versions TensorFlow versions prior to 2.11.0 TensorFlow versions 2.8.4, 2.9.3, and 2.10.1 are affected, but will be patched with a cherrypick commit. Description The function MakeGrapplerFunctionItem takes arguments that determine the sizes of inputs a...

PT-2022-26141

Name of the Vulnerable Software and Affected Versions TensorFlow versions prior to 2.11.0 TensorFlow versions 2.8.4 through 2.10.1 Description The issue is related to the function MakeGrapplerFunctionItem which takes arguments that determine the sizes of inputs and outputs. If the inputs given ar...

py-tensorflow -- denial of service vulnerability

Kang Hong Jin, Neophytos Christou, 刘力源 and Pattarakrit Rattankul report: Another instance of CVE-2022-35935, where SobolSample is vulnerable to a denial of service via assumed scalar inputs, was found and fixed. Pattarakrit Rattankul reports: Another instance of CVE-2022-35991, where...

Denial of Service (DoS)

Overview Affected versions of this package are vulnerable to Denial of Service DoS when SparseFillEmptyRowsGrad is given empty inputs. Details Denial of Service DoS describes a family of attacks, all aimed at making a system inaccessible to its intended and legitimate users. Unlike other...

Out-of-Bounds

Overview Affected versions of this package are vulnerable to Out-of-Bounds in DynamicStitch due to missing validation when it receives a differing number of inputs, such as when it is called with an indices size 1 and a data size 2. Remediation Upgrade tensorflow-lite to version 2.12.0 or higher...

Always-Incorrect Control Flow Implementation

Overview Affected versions of this package are vulnerable to Always-Incorrect Control Flow Implementation when a numpy array is created with a shape such that one element is zero and the sum of others is a large number. Remediation Upgrade tensorflow-lite to version 2.12.0 or higher. References -...

Buffer Overflow

Overview Affected versions of this package are vulnerable to Buffer Overflow via tf.rawops.ImageProjectiveTransformV2 when a large output shape is given. Remediation Upgrade tensorflow-lite to version 2.12.0 or higher. References - GitHub Commit - Vulnerable Code Credit: Neophytos Christou from S...

Incorrect Calculation of Buffer Size

Overview Affected versions of this package are vulnerable to Incorrect Calculation of Buffer Size via tf.keras.losses.poisson which receives a ypred and ytrue that are passed through functor::mul in BinaryOp. If the resulting dimensions overflow an int32, TensorFlow will crash due to a size...

Out-of-bounds Read

Overview Affected versions of this package are vulnerable to Out-of-bounds Read when the BaseCandidateSamplerOp function receives a value in trueclasses larger than rangemax. Remediation Upgrade tensorflow-lite to version 2.12.0 or higher. References - GitHub Commit - Vulnerable Code Credit: Yu...