Arbitrary Code Execution

tensorflow is vulnerable to arbitrary code execution. An heap-based buffer overflow in the implementation of SparseFillEmptyRowsGrad allows an attacker to execute arbitrary code on the host OS since it is possible for reverseindexmapi to be an index outside of bounds of gradvalues...

Information Disclosure

tensorflow is vulnerable to information disclosure. Lack of validation of weights tensor against the data in the functions SparseCountSparseOutput and RaggedCountSparseOutput allows a user to pass fewer weights than the values for the tensors to read out of heap buffer boundary, potentially...

Heap Buffer Overflow

tensorflow is vulnerable to heap buffer overflow. Failure to ensure that only operators which accept optional inputs use the -1 special value and only for the tensors that they expect to be optional allows an attacker to write and read from outside the bounds of heap allocated arrays...

Denial Of Service (DoS)

tensorflow is vulnerable to denial of service DoS. The vulnerability exists when the eager runtime traverses all tensors in the output, and a tensor could be a null pointer when it is processed from tf.rawops.Switch...

Denial Of Service (DoS)

tensorflow is vulnerable to denial of service DoS. The vulnerability exists due to a memory leak in the status argument when validation failures. This is caused lack of validation of the list of strings that is passed into dlpack.todlpack...

Information Disclosure

tensorflow is vulnerable to information disclosure. The vulnerability exists due to the lack of validation of the datasplits argument in tf.rawops.StringNGrams, resulting in a memory leak...

Denial Of Service (DoS)

tensorflow is vulnerable to denial of service DoS. The vulnerability exists due to the lack of validation of input arguments in RaggedCountSparseOutput, allowing a split tensor with exactly one element, or an empty split tensor to cause a SIGABRT signal...

Out-of-bounds (OOB) Read/Write

tensorflow is vulnerable to out-of-bounds OOB read/write. Lack of proper handling of the last argument of Shard API, a function taking two arguments of the lambda allows an attacker to lead to a truncation of arguments from positive int64 values to negative int ones, resulting in a buffer out of...

Format String Attack

tensorflow is vulnerable to format string attacks. The vulnerability exists as the fill argument of tf.strings.asstring reaches a printf call without sanitization...

Arbitrary Code Execution

tensorflow is vulnerable to arbitrary code execution. The vulnerability exists through a memory corruption issue in the implementation of dlpack.todlpack, as the pybind11 does not restrict Python objects to be parsed and casted...

Denial Of Service (DoS)

tensorflow is vulnerable to denial of service DoS. The vulnerability exists through a null pointer dereference that occurs in c/eager/dlpack.cc through an invalid argument to dlpack.todlpack...

Google TensorFlow Buffer Overflow Vulnerability (CNVD-2020-54656)

Google TensorFlow is a suite of end-to-end open source platforms for machine learning from Google USA. A security vulnerability exists in TensorFlow Lite versions prior to 2.2.1, 2.3.1, which can be exploited by an attacker to perform out-of-bounds memory operations, resulting in memory leakage...

Google TensorFlow Buffer Overflow Vulnerability (CNVD-2020-54655)

Google TensorFlow is a suite of end-to-end open source platforms for machine learning from Google USA. A security vulnerability exists in tensorflow-lite versions prior to 1.15.4, 2.0.3, 2.1.2, 2.2.1, 2.3.1, and 2.3.1, which can be exploited by an attacker to write and read from outside the...

Google TensorFlow Input Validation Error Vulnerability

Google TensorFlow is a suite of end-to-end open source platforms for machine learning from Google USA. A security vulnerability exists in Tensorflow versions prior to 1.15.4, 2.0.3, 2.1.2, 2.2.1, 2.3.1, and 2.3.1, which stems from a lack of validation of the datasplits parameter of...

Google TensorFlow Input Validation Error Vulnerability (CNVD-2020-55182)

Google TensorFlow is a suite of end-to-end open source platforms for machine learning from Google USA. An input validation error vulnerability exists in Tensorflow versions prior to 1.15.4, 2.0.3, 2.1.2, 2.2.1, 2.3.1, and 2.3.1, which stems from the program's failure to adequately perform path...

Google TensorFlow Buffer Overflow Vulnerability (CNVD-2020-54782)

Google TensorFlow is a suite of end-to-end open source platforms for machine learning from Google USA. A security vulnerability exists in Tensorflow SparseFillEmptyRowsGrad versions prior to 1.15.4, 2.0.3, 2.1.2, 2.2.1, 2.3.1, and 2.3.1, which arises from a networked system or product that perfor...

Tensorflow Data Validation Vulnerability

Google TensorFlow is a suite of end-to-end open source platforms for machine learning from Google USA. A security vulnerability exists in Tensorflow version 2.3.0 that stems from the inability of the SparseCountSparseOutput and RaggedCountSparseOutput implementations to verify that the weights...

Google TensorFlow code issue vulnerability (CNVD-2020-54781)

Google TensorFlow is a suite of end-to-end open source platforms for machine learning from Google USA. A security vulnerability exists in Tensorflow versions prior to 2.2.1, and in version 2.3.1, which stems from the pybind11 bonding code that assumes that the arguments are tensors, which can be...

Google TensorFlow Input Validation Error Vulnerability (CNVD-2020-54472)

Google TensorFlow is a suite of end-to-end open source platforms for machine learning from Google USA. A security vulnerability exists in Tensorflow versions prior to 1.15.4, 2.0.3, 2.1.2, 2.2.1, 2.3.1, which can be exploited by an attacker to trigger a format string vulnerability...

Google TensorFlow code issue vulnerability (CNVD-2020-54473)

Google TensorFlow is a suite of end-to-end open source platforms for machine learning from Google USA. A security vulnerability exists in TensorFlow eager mode versions prior to 1.15.4, 2.0.3, 2.1.2, 2.2.1, 2.3.1, and 2.3.1, which can be exploited by attackers to cause segmentation errors...