PT-2026-59740

Impact python import os os.environ'TF ENABLE ONEDNN OPTS' = '0' import tensorflow as tf printtf. version with tf.device"CPU": ksize = 1, 40, 128, 1 strides = 1, 128, 128, 30 padding = "SAME" data format = "NHWC" orig input shape = 11, 9, 78, 9 grad = tf.saturate casttf.random.uniform16, 16, 16, 1...

PT-2026-59850

Impact When tf.quantization.fake quant with min max vars gradient receives input min or max that is nonscalar, it gives a CHECK fail that can trigger a denial of service attack. python import tensorflow as tf import numpy as np arg 0=tf.constantvalue=np.random.randomsize=2, 2, shape=2, 2,...

PT-2026-59909

Impact When mlir::tfg::GraphDefImporter::ConvertNodeDef tries to convert NodeDefs without an op name, it crashes. cpp Status GraphDefImporter::ConvertNodeDefOpBuilder &builder, ConversionState &s, const NodeDef &node VLOG4 op def; else auto it = function op defs .findnode.op; if it == function op...

PT-2026-59914

Impact If the parameter indices for DynamicStitch does not match the shape of the parameter data, it can trigger an stack OOB read. python import tensorflow as tf func = tf.raw ops.DynamicStitch para='indices': 0xdeadbeef, 405, 519, 758, 1015, 'data': 110.27793884277344, 120.29475402832031,...

PT-2026-59935

Impact Integer overflow occurs when 2^31 = num frames height width channels 2^32, for example Full HD screencast of at least 346 frames. python import urllib.request dat =...

PT-2026-59803

Impact When AudioSummaryV2 receives an input sample rate with more than one element, it gives a CHECK fails that can be used to trigger a denial of service attack. python import tensorflow as tf arg 0='' arg 1=tf.random.uniformshape=1,1, dtype=tf.float32, maxval=None arg...

PT-2026-59806

Impact NPE in RandomShuffle with XLA enable python import tensorflow as tf func = tf.raw ops.RandomShuffle para = 'value': 1e+20, 'seed': -4294967297, 'seed2': -2147483649 @tf.functionjit compile=True def test: y = funcpara return y test Patches We have patched the issue in GitHub commit...

PT-2026-59980

Impact If the stride and window size are not positive for tf.raw ops.AvgPoolGrad, it can give an FPE. python import tensorflow as tf import numpy as np @tf.functionjit compile=True def test: y = tf.raw ops.AvgPoolGradorig input shape=1,0,0,0, grad=0.39117979, ksize=1,0,0,0, strides=1,0,0,0,...

PT-2026-59802

Impact An input sparse matrix that is not a matrix with a shape with rank 0 will trigger a CHECK fail in tf.raw ops.SparseMatrixNNZ. python import tensorflow as tf tf.raw ops.SparseMatrixNNZsparse matrix= Patches We have patched the issue in GitHub commit f856d02e5322821aad155dad9b3acab1e9f5d693...

PT-2026-59901

Impact Out of bounds read in GRUBlockCellGrad python func = tf.raw ops.GRUBlockCellGrad para = 'x': 21.1, 156.2, 83.3, 115.4, 'h prev': array136.5, 136.6, 'w ru': array26.7, 0.8, 47.9, 26.1, 26.2, 26.3, 'w c': array 0.4, 31.5, 0.6, 'b ru': array0.1, 0.2 , dtype=float32, 'b c': 0x41414141, 'r':...

PT-2026-59844

Impact The implementation of Conv2DBackpropInput requires input sizes to be 4-dimensional. Otherwise, it gives a CHECK failure which can be used to trigger a denial of service attack: python import tensorflow as tf strides = 1, 1, 1, 1 padding = "SAME" use cudnn on gpu = True explicit paddings =...

PT-2026-59974

Impact The implementation of Conv2DBackpropInput requires input sizes to be 4-dimensional. Otherwise, it gives a CHECK failure which can be used to trigger a denial of service attack: python import tensorflow as tf strides = 1, 1, 1, 1 padding = "SAME" use cudnn on gpu = True explicit paddings =...

PT-2026-59772

Impact When TensorListFromTensor receives an element shape of a rank greater than one, it gives a CHECK fail that can trigger a denial of service attack. python import tensorflow as tf arg 0=tf.random.uniformshape=6, 6, 2, dtype=tf.bfloat16, maxval=None arg 1=tf.random.uniformshape=6, 9, 1, 3,...

PT-2026-59849

Impact If LowerBound or UpperBound is given an emptysorted inputs input, it results in a nullptr dereference, leading to a segfault that can be used to trigger a denial of service attack. python import tensorflow as tf out type = tf.int32 sorted inputs = tf.constant, shape=10,0, dtype=tf.float32...

PT-2026-59883

Impact When tf.raw ops.ResizeNearestNeighborGrad is given a large size input, it overflows. import tensorflow as tf align corners = True half pixel centers = False grads = tf.constant1, shape=1,8,16,3, dtype=tf.float16 size = tf.constant1879048192,1879048192, shape=2, dtype=tf.int32 tf.raw...

PT-2026-59866

Impact tf.raw ops.DynamicStitch specifies input sizes when it is registered. cpp REGISTER OP"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 a...

PT-2026-59922

Impact When tf.linalg.matrix rank receives an empty input a, the GPU kernel gives a CHECK fail that can be used to trigger a denial of service attack. python import tensorflow as tf a = tf.constant, shape=0, 1, 1, dtype=tf.float32 tf.linalg.matrix ranka=a Patches We have patched the issue in GitH...

PT-2026-59773

Impact When tf.linalg.matrix rank receives an empty input a, the GPU kernel gives a CHECK fail that can be used to trigger a denial of service attack. python import tensorflow as tf a = tf.constant, shape=0, 1, 1, dtype=tf.float32 tf.linalg.matrix ranka=a Patches We have patched the issue in GitH...

PT-2026-59890

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

PT-2026-59778

Impact The function MakeGrapplerFunctionItem takes arguments that determine the sizes of inputs and outputs. If the inputs given are greater than or equal to the sizes of the outputs, an out-of-bounds memory read or a crash is triggered. Patches We have patched the issue in GitHub commit...