Skip to content

Bug: finegrained-fp8 act_quant produces NaNs on zero blocks #681

@wz-ml

Description

@wz-ml

Heya, quick note that finegrained-fp8's act_quant kernel computes the per-block FP8 scale with s = max(abs(x)) / 448 with no eps, so we get s = 0 when x is a block of zeros, and the 0/0 division yields NaN.

I opened a PR for it earlier but didn't realize that it was contributor-only: #680

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions