This repository has been archived by the owner on Dec 18, 2023. It is now read-only.
Warn user against using torch tensors as arguments of random variables #1639
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
Bean Machine uses the hash value of the arguments of random variables to identify them, which means that foo(1) and foo(tensor(1)) are considered two different random variables. Further in PyTorch, tensors are hashed by memory address instead of by value, so we can have hash(tensor(1)) != hash(tensor(1)). Therefore, it’s not recommended to use tensors as indices of random variables.
In this change,
rv_identifier.py
we identify if tensors are used as arguments to RVs and warn the user against its use.rv_identifier_test.py
to check if the warning is triggered correctly when the user provides a tensor instead of a primitive argument.Differential Revision: D39169577