flashinfer.fp4_quantization.SfLayout

class flashinfer.fp4_quantization.SfLayout(value, names=<not given>, *values, module=None, qualname=None, type=None, start=1, boundary=None)

Layout of scale factors for NVFP4.

__init__(*args, **kwds)

Attributes

layout_128x4

layout_8x4

layout_linear