2 Commits

Author	SHA1	Message	Date
biondizzle	2b1fca6dae	CRITICAL FIX: runtime activation global scale to prevent E4M3 overflow ... The checkpoint's input_scale was designed for training-time FP8 quantization, not NVFP4 activation quantization. Using it as gsa causes x/gsa to exceed the E4M3 block scale maximum (448), leading to systematic magnitude loss in every projection. This accumulates over 61 layers, compressing the logit range and producing garbage tokens. Fix: compute gsa at runtime from actual activation magnitude: gsa = max(|x|) / (6.0 * 448.0) This ensures x/gsa ≤ 2688 (the maximum representable in E4M3 block scales). Applied to: Nvfp4Linear, Nvfp4GroupedLinear, Nvfp4MoE, Nvfp4SharedExpert, Router gate	2026-06-01 14:21:16 +00:00
biondizzle	3fb3c925af	Restructure: cutedsl/ -> dsv4/ with proper layering ... - Split bridge.py -> ops/quantize.py, ops/layouts.py, ops/gemm_runner.py - Renamed classes: CuTeDSLNvfp4Linear -> Nvfp4Linear, etc. - Moved kernel code to dsv4/kernels/ (gemm, attention, compressor, decode, cuda) - Moved PyTorch bridges to dsv4/ops/ - Moved nn.Module layers to dsv4layers/ - Moved reference implementations to dsv4/reference/ - Moved vendored CUTLASS code to vendored/ - Archived ~190 debug tests to tests/archive/ - Kept ~15 canonical tests in tests/unit/ - Updated all import paths - Added stubs for future components (model/, cache/, loader/) - Updated pyproject.toml: dsv4-inference package name	2026-05-21 17:30:44 +00:00