Skip to content

[C++] Make float16 part of FloatingPointTypes() #51311

Description

@fornwall

Describe the enhancement requested

FloatingPointTypes() currently contains only float32() and float64(), so float16 is also absent from NumericTypes() and PrimitiveTypes().

As a result, every kernel registration loop and test suite that iterates over those lists skips float16, and support for it has to be added type by type instead. Several kernel families explicitly reject HalfFloatType with Status::NotImplemented today.

float16 should probably be a member of FloatingPointTypes(), consistent with the other floating point types.

Component(s)

C++

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions