Files
nvfp4-megamoe-kernel/single_shot_inference.py
biondizzle 53178d2536 Add emergency RMSNorm after residuals (missing mHC fallback)
Without mHC, values explode to 761K after first layer.
Added per-residual RMSNorm + BF16 clamp to keep values bounded.
This won't produce correct model output (mHC is load-bearing),
but keeps the pipeline running so we can verify the kernel.
2026-05-30 22:56:16 +00:00

17 KiB