biondizzle 94b30dc2bc revert: block_n/4 was correct (SwiGLU halving × FP4 packing)

2026-05-12 15:04:23 +00:00

..

heuristics

NVFP4: fix SF pipeline — 2 K-cols per BLOCK_K for group=16

2026-05-12 08:08:17 +00:00

impls

revert: block_n/4 was correct (SwiGLU halving × FP4 packing)

2026-05-12 15:04:23 +00:00