[Quantization] Saturate NVFP4 export FP8 scale cast to avoid NaN#1397

cjluo-nv merged 1 commit into

mainNVIDIA/Model-Optimizer:mainfrom

chenjiel/fix-nvfp4-fp8-scale-cast-nanNVIDIA/Model-Optimizer:chenjiel/fix-nvfp4-fp8-scale-cast-nanCopy head branch name to clipboard