biondizzle e8a7a9256f fix: convert uint8 checkpoint weights to float4_e2m1fn_x2 for CuTeDSL GEMM ...

The CuTeDSL kernel expects float4_e2m1fn_x2 dtype for FP4 weight tensors,
but checkpoint weights from safetensors are loaded as uint8. The uint8 and
float4_e2m1fn_x2 have the same byte representation, so .view() is safe.

Fixed in:
- Nvfp4Linear.finalize_weights()
- Nvfp4SharedExpert.finalize_weights()
- Nvfp4MoE._ensure_stacked() (both stacked and legacy paths)

2026-06-01 00:18:34 +00:00

..

cache

E1: Wire LayerCacheHandle gather methods + CUDA gather kernels

2026-05-30 21:09:21 +00:00

kernels

fix: W_gate is (H, E) but F.linear expects (E, H), transpose before linear

2026-05-31 23:55:16 +00:00

layers

fix: convert uint8 checkpoint weights to float4_e2m1fn_x2 for CuTeDSL GEMM

2026-06-01 00:18:34 +00:00

loader

Restructure: cutedsl/ -> dsv4/ with proper layering

2026-05-21 17:30:44 +00:00

model

Fix remaining mHC API references: layer_compare.py, layer.py comment

2026-05-31 18:38:34 +00:00

ops

fix: import SF_VEC_SIZE from quantize in gemm_runner (was NameError)

2026-06-01 00:04:48 +00:00

reference

Restructure: cutedsl/ -> dsv4/ with proper layering

2026-05-21 17:30:44 +00:00

__init__.py

Restructure: cutedsl/ -> dsv4/ with proper layering

2026-05-21 17:30:44 +00:00