biondizzle e671780008 fix: transpose checkpoint weights before make_b_k_major in Nvfp4Linear/SharedExpert ...

Critical bug: checkpoint weights are (N_packed, K_packed) N-major format,
but make_b_k_major expects (E, K_packed, N_packed) input. Without the
permute, the K and N dimensions are swapped, producing garbage output
with wrong dimensions (e.g., q_a output was 3584 instead of 1536).

Also fix scale assembly: checkpoint scales are (N, K_sf) which should
use assemble_raw_scales_2d3d_3d_side (no transpose), not
assemble_scales_3d_side (which incorrectly transposes K_sf↔N).

2026-06-01 00:30:37 +00:00

7.3 KiB

Raw Blame History

View Raw