feat: Enforce ml_dtypes.bfloat16 for BF16 I/O in Python client#897
Open
yinggeh wants to merge 1 commit into
Open
feat: Enforce ml_dtypes.bfloat16 for BF16 I/O in Python client#897yinggeh wants to merge 1 commit into
ml_dtypes.bfloat16 for BF16 I/O in Python client#897yinggeh wants to merge 1 commit into
Conversation
yinggeh
changed the title
yinggeh
changed the title
yinggeh
changed the title
yinggeh
changed the title
ml_dtypes.bfloat16 for BF16 I/O in Python client
whoisj
requested changes
May 15, 2026
| dtype = np_to_triton_dtype(input_tensor.dtype) | ||
| if self._input.datatype != dtype: | ||
| error_message = ( | ||
| "got unexpected datatype {} from numpy array, expected {}.".format( |
Contributor
There was a problem hiding this comment.
why use an interpolated string instead?
f"got unexpected datatype {dtype} from numpy array, expected {self._input.datatype}."
| dtype = np_to_triton_dtype(input_tensor.dtype) | ||
| if self._datatype != dtype: | ||
| error_message = ( | ||
| "got unexpected datatype {} from numpy array, expected {}.".format( |
Contributor
There was a problem hiding this comment.
again, why not an interpolated string?
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
NumPy has no native BF16. Currently, casting FP32 <=> BF16 magnifies accuracy loss. Enforce using
ml_dtypes.bfloat16for native BF16 support.Example
For 0.01 + 0.01, BF16 <=> FP32 will lose precision during truncating, making the output value biases from the true result 0.02.