biondizzle 2ef71dc21a fix: B tensor K-major strides, scale_b axis swap ...

Two fixes:
1. B tensor: permute(0,2,1).contiguous().permute(0,2,1) gives K-major
   stride (16384,1,128) matching reference
2. scale_b: transpose to (N, K_sf) before swizzling — reference uses
   (intermediate, hidden//16) not (hidden//16, intermediate)

2026-05-16 03:04:31 +00:00

..

layertest.py

test: streamline layertest — kernel vs BF16 ref only, exit on fail

2026-05-16 02:29:41 +00:00

requirements.txt

test: add standalone layer 0 comparison test (no vLLM, no Docker)

2026-05-16 02:13:18 +00:00

run_test.sh

fix: use setup.py install for CUTLASS extension build

2026-05-16 02:21:17 +00:00

test_b_layout.py

cleanup: move useful tests to tests/, nuke stale debug tests

2026-05-16 02:14:37 +00:00

test_cutedsl.py

fix: B tensor K-major strides, scale_b axis swap

2026-05-16 03:04:31 +00:00

test_quick_rand.py

cleanup: move useful tests to tests/, nuke stale debug tests

2026-05-16 02:14:37 +00:00

test_uniform_fp4.py

cleanup: move useful tests to tests/, nuke stale debug tests

2026-05-16 02:14:37 +00:00