biondizzle 70c0618361 fix: allocate token_indices before CuTeDSL JIT compilation ...

CuTeDSL's cute.compile appears to corrupt GPU memory state,
causing torch.arange to produce zero-filled tensors when allocated
after the JIT compilation. Moving token_indices allocation before
the weight stacking operations fixes the corruption.

2026-05-17 08:20:41 +00:00

..

patches

fix: set _l1_activation_global_scale (with underscore) — attribute name mismatch

2026-05-17 03:35:20 +00:00

nvfp4_cutedsl.py

fix: allocate token_indices before CuTeDSL JIT compilation

2026-05-17 08:20:41 +00:00