I took remote ML class with LD. Company paid

來源: RomaVacation 於 2025-02-14 08:48:47 [] [舊帖] [給我悄悄話] 閱讀數 : (232 bytes)
回答: 落子無悔 由 凊荷 於 2025-02-14 08:23:07

I don't like the blackbox approach. You unfreeze certain layers in neural networks and results become better. 

Too slow to run some on CPU. Had to pay google for TPU and GPU access. 

 

請您先登陸，再發跟帖！