Fix momo crash when loss value is a Python float #1502
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Outdated suggestions cannot be applied.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.
Fixes #1500
The momo and momo_adam update functions declared the value parameter as Optional[jax.Array] but never enforced that type. When calling value.astype(state.barf.dtype), passing a Python float or NumPy scalar would raise AttributeError since these types don't have an astype method.
This fix uses jnp.asarray(value, dtype=state.barf.dtype) instead of value.astype(state.barf.dtype) to handle Python floats, NumPy scalars, and JAX arrays uniformly.