feat(speculative): DFlash-family loss terms in config; one teacher with an opt-in softcap - #2993

feat(speculative): DFlash-family loss terms in config; one teacher with an opt-in softcap#2993

h-guo18 wants to merge 4 commits into

haoguo/dflash-shared-forwardNVIDIA/Model-Optimizer:haoguo/dflash-shared-forwardfrom

haoguo/dflash-loss-terms-configNVIDIA/Model-Optimizer:haoguo/dflash-loss-terms-configCopy head branch name to clipboard