biondizzle fcd6de0a60 test: simplify SF fill to avoid shape mismatch

2026-05-12 15:13:16 +00:00

..

deepseek_v4.py

fix: use torch.int8 for packed FP4 tensors (kPackedFP4=kInt8, not uint8)

2026-05-12 12:23:43 +00:00

staging_kernel.py

fix staging kernel packed_k_mask double-count

2026-05-12 08:08:24 +00:00

test_nvfp4_mega_moe.py

test: simplify SF fill to avoid shape mismatch

2026-05-12 15:13:16 +00:00