IMPORTANT: To view this page as Markdown, append `.md` to the URL (e.g. /get-started.md). For the complete documentation index, see llms.txt.

Skip to main content

For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).

Mojo module

mxfp6_dequant

MXFP6 dequantization kernel.

Converts packed MXFP6 weights (uint8, four 6-bit values per three bytes) with E8M0 block scales into bfloat16 or float8_e4m3fn. NVIDIA and AMD only.

Scales are 2D [N, K/SF_VECTOR_SIZE], each covering SF_VECTOR_SIZE (32) consecutive elements.

`comptime` values

`BYTES_PER_THREAD`

comptime BYTES_PER_THREAD = 24

`ELEMENTS_PER_THREAD`

comptime ELEMENTS_PER_THREAD = MXFP6_SF_VECTOR_SIZE

Functions

dequant_mxfp6: Dequantizes MXFP6 packed weights to BF16 or FP8.

comptime values
- BYTES_PER_THREAD
- ELEMENTS_PER_THREAD
Functions