[2/4] perf(speculative): deliver training batches out of order; compile + FlexAttention in the DFlash2 recipe - #3001

[2/4] perf(speculative): deliver training batches out of order; compile + FlexAttention in the DFlash2 recipe#3001

h-guo18 wants to merge 3 commits into

haoguo/dflash-rope-fp32NVIDIA/Model-Optimizer:haoguo/dflash-rope-fp32from

haoguo/spec-train-inorder-compileNVIDIA/Model-Optimizer:haoguo/spec-train-inorder-compileCopy head branch name to clipboard