Luka Govedič 0ebf4e969b Add checks for hidden size and input stride divisibility by 4, fix unit test ...

Signed-off-by: Luka Govedič <lgovedic@redhat.com>

2026-03-10 11:54:08 -04:00

..

fused_layernorm_dynamic_per_token_quant.cu

Add checks for hidden size and input stride divisibility by 4, fix unit test

2026-03-10 11:54:08 -04:00

layernorm_utils.cuh

Fix kernel offset calculation

2026-03-10 11:53:56 -04:00

quant_conversions.cuh

[Refactor] Refactor FP8 & INT8 Quant Folder inside w8a8 (#25293)

2025-10-08 10:20:48 -04:00