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
We talked and came to the conclusion that we would like to add static type checking as a GHA to this repository. This would depend on #2 due to some missing type annotations. I ran some type checking efforts a while back, and while there are some errors, most of the files are fine. Apart from having to generate some stubs, it would be a fairly easy thing to implement. I'd propose using mypy since it is the most commonly used framework for type checking.
The text was updated successfully, but these errors were encountered:
Depending on how strictly you lint your code, there can be an typing error when overriding Transformation.apply(). Since the abstract class in QONNX has no typing annotations, its return type is assigned None and thus clashes with the actual type used in FINN: tuple[ModelWrapper, bool]. We could either submit a PR to QONNX (probably not a good idea, since other projects aside from FINN may return different objects) or generate stubs or force the linter to ignore the line.
We talked and came to the conclusion that we would like to add static type checking as a GHA to this repository. This would depend on #2 due to some missing type annotations. I ran some type checking efforts a while back, and while there are some errors, most of the files are fine. Apart from having to generate some stubs, it would be a fairly easy thing to implement. I'd propose using mypy since it is the most commonly used framework for type checking.
The text was updated successfully, but these errors were encountered: