Division by zero in TFLite's implementation of `OneHot`
tensorflow
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI |
|
|
tensorflow
| Product | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI |
|
|
| tensorflow-cpu PyPI |
|
|
| tensorflow-gpu PyPI |
|
|
TensorFlow is an end-to-end open source platform for machine learning. The implementation of the `OneHot` TFLite operator is vulnerable to a division by zero error(https://github.com/tensorflow/tensorflow/blob/f61c57bd425878be108ec787f4d96390579fb83e/tensorflow/lite/kernels/one_hot.cc#L68-L72). An attacker can craft a model such that at least one of the dimensions of `indices` would be 0. In turn, the `prefix_dim_size` value would become 0. The fix will be included in TensorFlow 2.5.0. We will also cherrypick this commit on TensorFlow 2.4.2, TensorFlow 2.3.3, TensorFlow 2.2.3 and TensorFlow 2.1.4, as these are also affected and still in supported range.
From the CVE record · GitHub_M
The TFLite OneHot operator can divide by zero when it processes a crafted model with a zero-sized dimension in `indices`. An attacker who can supply a model for processing can trigger this error; the input does not specify network access, authentication requirements, or additional user interaction.
Check deployments using affected versions of the PyPI `tensorflow`, `tensorflow-cpu`, or `tensorflow-gpu` packages: below 2.1.4, 2.2.0–2.2.2, 2.3.0–2.3.2, or 2.4.0–2.4.1.