Enable to assign different parameters dtype during training #1037
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jialingt
changed the title
jialingt
changed the title
apghml
approved these changes
Mar 7, 2025
markblee
reviewed
Mar 7, 2025
| def __call__(self, params: Union[Nested[Tensor], Nested[TensorSpec]]) -> Nested[T]: | ||
| """This protocol requires a callable that accepts either a nested Tensor or | ||
| a nested TensorSpec as input and returns a processed value for each parameter. | ||
| Args: |
There was a problem hiding this comment.
Looks like some docstring format got lost from copy/paste?
There was a problem hiding this comment.
Oh you are right! Good catch.
Added all spaces back.
markblee
approved these changes
Mar 7, 2025
|
Hi, after this PR, CI is broken. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
This PR enables
train_dtypeto accept bothjnp.dypeand alsoPerParamFn[jnp.dtype]:jnp.dtype, where both float inputs and model parameters will be cast to this dtype.ConfigOr[PerParamFn[jnp.dtype]], allowing different dtypes to be applied to different parameters during training.