Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Exclude certain operators from injection #14

Open
nniranjhana opened this issue Nov 2, 2019 · 1 comment
Open

Exclude certain operators from injection #14

nniranjhana opened this issue Nov 2, 2019 · 1 comment
Labels
enhancement New feature or request

Comments

@nniranjhana
Copy link
Contributor

Injecting operators like Shape should be excluded. Accuracy reported might be wrong in these cases, as number of elements changes.

Also, curate a list of operators which just get values from the tf object, and/or not involved in computation and exclude them for fault injection.

Choosing "ALL" in the Ops configuration file will then work better.

@nniranjhana nniranjhana added the enhancement New feature or request label Nov 2, 2019
@nniranjhana
Copy link
Contributor Author

nniranjhana commented Nov 4, 2019

This is a list of operators I think we should exclude.
If you differ from this list, please unmark and state why. Alternatively, do add more too:

  • tf.shape
  • tf.rank
  • tf.one_hot

Should operators like tf.expand_dims and tf.reshape be excluded? Their outputs are not a dimension, but tensors, so it would be sane to inject faults into them. But does injecting faults into these operators make sense?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request
Projects
None yet
Development

No branches or pull requests

1 participant