biondizzle/nvfp4-megamoe-kernel
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
0d1cd1e21675d1947eccfe7282e4fc0d60feee23
nvfp4-megamoe-kernel/dsv4
History
biondizzle 0d1cd1e216 P4: Add QuantizedActivation + Nvfp4Linear.run_from_quantized 
- QuantizedActivation: carries (x_fp4, x_sf, gsa) for skip-quantize path
- Nvfp4Linear.run_from_quantized(): runs GEMM with pre-quantized input
- Enables fused RMSNorm+quantize to feed directly into all downstream
  linears (q_a, kv, o_proj, etc.) without re-quantizing
2026-06-02 16:37:38 +00:00
..
cache
E1: Wire LayerCacheHandle gather methods + CUDA gather kernels
2026-05-30 21:09:21 +00:00
kernels
P4: Fix fused RMSNorm kernel — match quantize_nvfp4.cu encoding
2026-06-02 16:28:44 +00:00
layers
P4: Add QuantizedActivation + Nvfp4Linear.run_from_quantized
2026-06-02 16:37:38 +00:00
loader
Restructure: cutedsl/ -> dsv4/ with proper layering
2026-05-21 17:30:44 +00:00
model
P0 COMPLETE: Eliminate ALL .item() CPU-GPU syncs from NVFP4 activation path
2026-06-01 21:05:03 +00:00
ops
P4: Add QuantizedActivation + Nvfp4Linear.run_from_quantized
2026-06-02 16:37:38 +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