Optimize quantized MoE and block-scaled FP8 CUDA projections - #33463

Optimize quantized MoE and block-scaled FP8 CUDA projections#33463

Akshay Sonawane (apsonawane) wants to merge 4 commits into

asonawane/ported-ops-sparse-attention-indexermicrosoft/onnxruntime:asonawane/ported-ops-sparse-attention-indexerfrom

asonawane/ported-ops-quantized-moe-fp8microsoft/onnxruntime:asonawane/ported-ops-quantized-moe-fp8Copy head branch name to clipboard