You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
First of all, I am a fan of the Tensorflow framework, so I was very happy when I found that ModelScan supports security detection of Tensorflow models(Protocol Buffer format). This is great for me.
However, I carefully read the spec and source code of ModelScan, and I found that ModelScan's support for the Tensorflow framework seems to be a bit limited. It seems to only support the API detection of ReadFile, WriteFile, Lambda Layer and tf.io.matching_files.
ModelScan's support for Pytorch is very beautiful and elegant, so I am curious if ModelScan has further plans to improve its support for Tensorflow? Or are there other API calls in Tensorflow that will cause Model Serialization Attack?
The text was updated successfully, but these errors were encountered:
First of all, I am a fan of the Tensorflow framework, so I was very happy when I found that ModelScan supports security detection of Tensorflow models(Protocol Buffer format). This is great for me.
However, I carefully read the spec and source code of ModelScan, and I found that ModelScan's support for the Tensorflow framework seems to be a bit limited. It seems to only support the API detection of ReadFile, WriteFile, Lambda Layer and tf.io.matching_files.
ModelScan's support for Pytorch is very beautiful and elegant, so I am curious if ModelScan has further plans to improve its support for Tensorflow? Or are there other API calls in Tensorflow that will cause Model Serialization Attack?
The text was updated successfully, but these errors were encountered: