CVE-2021-37679

TensorFlow CVE-2021-37679 concerns a vulnerability in nested tf.map_fn with RaggedTensor inputs. The root cause is in the conversion from a Variant tensor to a RaggedTensor: the implementation does not verify that all inner shapes match, which can produce extra dimensions and allow leakage of hea...

CVE-2022-29206

CVE-2022-29206 involves TensorFlow’s tf.raw_ops.SparseTensorDenseAdd, where input argument validation is insufficient, causing a reference to a nullptr during kernel execution and resulting in undefined behavior. Affected releases include TensorFlow versions prior to 2.9.0, and also 2.8.1, 2.7.2,...

CVE-2021-37673

TensorFlow maps-stage DoS (CVE-2021-37673). The issue arises from tf.raw_ops.MapStage not validating non-empty key tensors, enabling a local attacker to trigger a denial-of-service via a CHECK-fail. Patches were committed and the fix is expected in TensorFlow 2.6.0, with backports to 2.5.1, 2.4.3...

CVE-2021-37689

CVE-2021-37689 : In TensorFlow’s TFLite MLIR optimizations, the L2NormalizeReduceAxis path dereferences a vector iterator without checking for elements, enabling a local attack to crash the process (DoS) via a crafted TFLite model. A patch was applied in the referenced GitHub commit (d6b57f461b39...

CVE-2022-23582

CVE-2022-23582 affects TensorFlow: a malicious SavedModel can trigger a denial of service via TensorByteSize CHECK failures caused by shape handling in TensorShape/PartialTensorShape (shape partials or large element counts). Root cause is TensorShape throwing on partial/large shapes; PartialTenso...

CVE-2022-23592

CVE-2022-23592 : TensorFlow’s type inference can trigger a heap out-of-bounds read when input_idx controls ix, risking access beyond node_t.args. This occurs because bounds checks run via DCHECK (no-op in production). The issue affects the TensorFlow release line noted as the only affected versio...

CVE-2021-29546

CVE-2021-29546 concerns TensorFlow’s QuantizedBiasAdd. The issue arises from the Eigen kernel implementation in quantization_utils.h, which divides by the number of elements of the smaller input without guarding against zero, causing integer division by zero undefined behavior. Public details in ...

CVE-2021-29552

TensorFlow vulnerability CVE-2021-29552 is a denial-of-service in UnsortedSegmentJoin caused by an out-of-bounds CHECK on an empty num_segments scalar. Affected TF versions include those in the 2.1.x–2.5.x range; the issue stems from treating num_segments as a valid scalar and failing when the te...

CVE-2021-29610

CVE-2021-29610: TensorFlow QuantizeAndDequantizeV2 accepts axis values

CVE-2022-23561

CVE-2022-23561 affects TensorFlow’s TFLite, enabling out-of-bounds writes by crafting a TFLite model that can corrupt the memory allocator’s linked list. This vulnerability allows an arbitrary write primitive under certain conditions as described in the CVE description. Affected details in connec...

CVE-2022-23562

TensorFlow vulnerability CVE-2022-23562 concerns the Range implementation: integer overflows in Range can cause undefined behavior or extremely large allocations. Public notes indicate a fix will be included in TensorFlow 2.8.0, with cherry-picks to affected supported releases (2.7.1, 2.6.3, 2.5....

CVE-2022-23569

TensorFlow/CVE-2022-23569: Denial‑of‑service via CHECK‑fails (assertion failures) across multiple ops in TensorFlow/TFLite, caused by invalid tensor shapes and related checks. Root cause: assertion failures in core paths; fixes released as patches implemented in GitHub commits. Remediation: upgra...

CVE-2022-29192

TensorFlow vulnerability CVE-2022-29192 affects tf.raw_ops.QuantizeAndDequantizeV4Grad; it does not fully validate input arguments, causing a CHECK failure that can trigger a denial of service. Affected versions are prior to 2.9.0, 2.8.1, 2.7.2, and 2.6.4. A patch exists in 2.9.0 and has been bac...

CVE-2022-29204

TensorFlow CVE-2022-29204 affects the tf.raw_ops.UnsortedSegmentJoin implementation in multiple pre-2.9.0 releases (2.9.0, 2.8.1, 2.7.2, 2.6.4). The issue arises from incomplete validation of input arguments, specifically num_segments, which is treated as a positive scalar and used to allocate th...

CVE-2021-35958

TensorFlow prior to 2.5.1/vulnerable until patched: TF 2.5.0 and earlier allow arbitrary file overwrite via a crafted archive when tf.keras.utils.get_file is used with extract=True. Root cause: get_file extracting an archive from untrusted sources can overwrite arbitrary files. Affected component...

CVE-2021-37675

CVE-2021-37675 : TensorFlow denial-of-service caused by a division-by-zero in the shape inference path of most convolution operators. Root cause: missing validations before divisions/modulo in the common_shape_fns.cc implementation. Patch is in commit 8a793b5d7f59e37ac7f3cd0954a750a2fe76bad4 and ...

CVE-2022-21737

The CVE-2022-21737 issue affects TensorFlow’s bincount path. The vulnerability arises from the implementation of DenseBincount and related *Bincount operations, where certain input argument constraints are not fully enforced during shape inference or kernel execution, leading to CHECK failures du...

CVE-2022-41896

CVE-2022-41896 affects TensorFlow: a crash occurs when ThreadUnsafeUnigramCandidateSampler receives filterbank_channel_count beyond the allowed max size. The issue has been patched in commit 39ec7eaf1428e90c37787e5b3fbd68ebd3c48860 and will be included in TensorFlow 2.11, with backport cherrypick...

CVE-2021-29540

TensorFlow CVE-2021-29540 describes a heap-based buffer overflow in Conv2DBackpropFilter caused by calculating the filter tensor size without validating it against filter_sizes. This mismatch can lead to out-of-bounds reads/writes when accessing the buffer. The issue affects multiple TensorFlow v...

CVE-2021-29545

CVE-2021-29545 concerns TensorFlow and describes a heap-based out-of-bounds issue in SparseTensorToCSRSparseMatrix caused by a double redirection when accessing csr_row_ptr via indices(i, 0) + 1. This can lead to a denial of service by writing outside heap data. The connected OSV/GHSA entries con...

CVE-2022-21731

Technical details are not publicly available in the provided Connected documents. The Initial Description mentions a TensorFlow shape-inference vulnerability but does not specify affected products/versions beyond general references. Monitor for updates and official advisories for precise impact a...

CVE-2022-35963

TensorFlow FractionalAvgPoolGrad is vulnerable due to incomplete validation of orig_input_tensor_shape, allowing an overflow that can trigger a CHECK failure and deny service. The issue is fixed in GitHub commit 03a659d7be9a1154fdf5eeac221e5950fec07dad and will be included in TensorFlow 2.10.0; t...

CVE-2022-35979

CVE-2022-35979 affects TensorFlow. When QuantizedRelu or QuantizedRelu6 receive nonscalar inputs for min_features or max_features, a segfault can be triggered, leading to a denial of service. A fix was committed (49b3824d83af706df0ad07e4e677d88659756d89) and will be included in TensorFlow 2.10.0....

CVE-2021-37663

CVE-2021-37663 arises from incomplete validation in tf.raw_ops.QuantizeV2 in TensorFlow. Affected: TensorFlow versions before 2.6.0 (patch committed 6da6620...). The issue can trigger undefined behavior via binding a null pointer or out-of-bounds access; if axis is provided, min_range/max_range l...

CVE-2021-37674

Summary (CVE-2021-37674) TensorFlow’s tf.raw_ops.MaxPoolGrad had insufficient validation for orig_input/orig_output, enabling a local attacker to trigger a denial-of-service via a segmentation fault. The issue is tied to CVE-2021-29579 and is addressed by patch 136b51f10903e044308cf77117c0ed98713...

CVE-2022-23563

TensorFlow (CVE-2022-23563) describes a TOCTOU race caused by tempfile.mktemp usage, where a temporary file could be created by another process between the check and the actual creation. Several connected sources confirm this insecure temporary-file pattern and note that the fix replaces mktemp w...

CVE-2022-29205

TensorFlow CVE-2022-29205 affects the TensorFlow project where calling tf.compat.v1.* ops that do not yet support quantized types can dereference a null value in ParseDimensionValue, leading to a segfault and potential DoS. Public details specify affected versions are prior to 2.9.0, 2.8.1, 2.7.2...

CVE-2022-35959

TensorFlow’s CVE-2022-35959 is a denial-of-service flaw caused by insufficient validation of orig_input_shape in AvgPool3DGradOp. The overflow triggers a CHECK failure, with impact limited to environments using affected builds. The issue has been patched in commit 9178ac9d6389bdc54638ab913ea0e419...

CVE-2022-41901

CVE-2022-41901 affects TensorFlow. The issue is a CHECK fail in tf.raw_ops.SparseMatrixNNZ triggered when input tensor is not a rank-0 matrix, which can cause a crash (denial of service). Fixed in GitHub commit f856d02e5322821aad155dad9b3acab1e9f5d693; the fix will be included in TensorFlow 2.11 ...

CVE-2021-29560

TensorFlow RaggedTensorToTensor heap-based overflow vulnerability (CVE-2021-29560) arises when the code uses the same index to access two arrays in parallel during ragged tensor to tensor conversion. An attacker-controlled input can trigger a heap OOB access when parent_output_index is shorter th...

CVE-2021-37670

TensorFlow vulnerability CVE-2021-37670 involves a heap-based out-of-bounds read in tf.raw_ops.UpperBound (and LowerBound) caused by missing rank validation of the sorted_input argument. A patch was committed (commit 42459e4273c2e47a3232cc16c4f4fff3b3a35c38) and the fix is slated for TensorFlow 2...

CVE-2021-37684

The CVE-2021-37684 issue affects TensorFlow/TFLite pooling implementations and is caused by division by zero due to missing checks for divisors. A patch was committed (dfa22b348b70bb89d6d6ec0ff53973bacb4f4695) and the fix is planned for TensorFlow 2.6.0, with cherry-picks to 2.5.1, 2.4.3, and 2.3...

CVE-2022-21725

TensorFlow CVE-2022-21725 describes a vulnerability in the cost estimator for some convolution operations where the stride is not properly validated, enabling a division-by-zero scenario. The issue is caused by the estimator failing to ensure the stride argument is strictly positive. The fix is t...

CVE-2022-23584

TensorFlow PNG decoding contains a use-after-free in png::CommonFreeDecode(&decode) where, after the call, decode.width and decode.height are in an unspecified state. This TF vulnerability (CVE-2022-23584) affects the TensorFlow PNG decode path and can lead to memory corruption or crashes when de...

CVE-2022-29203

CVE-2022-29203 describes an integer overflow in TensorFlow’s tf.raw_ops.SpaceToBatchND across backends (XLA and handwritten kernels) that can cause a denial of service via a CHECK failure when allocating the output tensor. Affected versions are prior to 2.9.0, 2.8.1, 2.7.2, and 2.6.4. The documen...

CVE-2022-29211

CVE-2022-29211 (TensorFlow) : The CPU implementation of tf.histogram_fixed_width crashes when the values array contains NaN. Specifically, casting NaN divisions to int32 can trigger a crash, affecting TensorFlow versions prior to 2.9.0, 2.8.1, 2.7.2, and 2.6.4. The issue is CPU-only and arises fr...

CVE-2021-37691

CVE-2021-37691 affects TensorFlow (Lite) where a crafted TFLite model can trigger a division-by-zero error in the LSH projection implementation. The issue is mitigated by a patch in the GitHub commit 0575b640091680cfb70f4dd93e70658de43b94f9, with the fix slated for inclusion in TensorFlow 2.6.0 a...

CVE-2022-23567

CVE-2022-23567 concerns TensorFlow: integer overflows in SparseCwise ops can trigger large allocations (OOM) or CHECK failures during TensorShape construction due to missing input-shape validation. The vulnerability affects TensorFlow releases prior to the fixed version and is acknowledged with a...

CVE-2022-23585

CVE-2022-23585 refers to a memory-leak vulnerability in TensorFlow's PNG decoding path. After calling png::CommonInitDecode(..., &decode), allocated buffers may remain if an error path triggers OP_REQUIRES, and are not freed before function termination, creating a potential leak. The issue affect...

CVE-2022-29194

CVE-2022-29194 affects TensorFlow where the implementation of tf.raw_ops.DeleteSessionTensor does not fully validate its input arguments, leading to a CHECK failure that can trigger a denial of service. The issue is fixed in versions 2.9.0, 2.8.1, 2.7.2, and 2.6.4, which contain patches. Connecte...

CVE-2022-35993

TensorFlow CVE-2022-35993 concerns a CHECK failure in SetSize when set_shape is not 1D, allowing a denial-of-service condition. The issue is attributed to SetSize input validation and is mitigated by a patch in commit cf70b79d2662c0d3c6af74583641e345fc939467, with fixes planned for TensorFlow 2.1...

CVE-2022-36004

CVE-2022-36004 affects TensorFlow. When using tf.random.gamma with large input shapes and rates, a CHECK failure can cause a denial of service. The issue was patched in commit 552bfced6ce4809db5f3ca305f60ff80dd40c5a3 and the fix will be included in TensorFlow 2.10.0; it will be cherry-picked to T...

CVE-2022-41890

CVE-2022-41890 (TensorFlow) : A bug in BCast::ToShape can crash TensorFlow when given input larger than int32, despite intended int64 support. A GitHub commit (8310bf8dd188ff780e7fc53245058215a05bdbe5) patches the issue; the fix will be in TensorFlow 2.11 and will be cherry-picked to 2.10.1, 2.9....

CVE-2020-15266

In TensorFlow, CVE-2020-15266 affects the tf.image.crop_and_resize path: when boxes contains a very large value in versions before 2.4.0, the CPU kernel reads it as a C++ NaN and causes undefined behavior leading to a segmentation fault. The issue is fixed in commit eccb7ec454e6617738554a255d77f0...

CVE-2021-29520

TensorFlow CVE-2021-29520 concerns a heap buffer overflow in Conv3DBackprop* due to missing validation that assumes input, filter_sizes, and out_backprop have identical shapes. Multiple sources (OSV entries and GHSA advisory) corroborate the issue and patch lineage. The vulnerability affects Conv...

CVE-2021-29544

TensorFlow vulnerability CVE-2021-29544 (QuantizeAndDequantizeV4Grad) is triggered by a rank-validation CHECK-fail in tf.raw_ops.QuantizeAndDequantizeV4Grad, caused by inputs not being validated before passing to QuantizeAndDequantizePerChannelGradientImpl. The issue leads to denial of service vi...

CVE-2021-29577

CVE-2021-29577 describes a heap-buffer-overflow in TensorFlow’s tf.raw_ops.AvgPool3DGrad caused by an assumption that orig_input_shape and grad have similar first/last dimensions without proper validation. Public sources (NVD, OSV, GHSA) confirm the issue and its impact as a memory corruption vul...

CVE-2021-29581

CVE-2021-29581 describes a denial-of-service vulnerability in TensorFlow’s tf.raw_ops.CTCBeamSearchDecoder, where lack of input validation on empty tensors allows a local attacker to trigger a segfault by reading from a null buffer. The issue affects TensorFlow versions in scope (2.x) and is addr...

CVE-2021-29582

TensorFlow CVE-2021-29582 describes a heap-out-of-bounds read in tf.raw_ops.Dequantize caused by a shape mismatch between min_range and max_range tensors. The vulnerability allows a local attacker to trigger reads outside heap data. The remediation is a patch, with fixes targeted for TensorFlow 2...

CVE-2022-29195

TensorFlow CVE-2022-29195 concerns a denial-of-service via missing validation in tf.raw_ops.StagePeek. Affected are versions before patches: 2.9.0, 2.8.1, 2.7.2, and 2.6.4, which patch the issue. The vulnerability stems from StagePeek assuming index is a scalar without validating input, leading t...