Thom Wolf (@Thom_Wolf)· @Thom_Wolf · X·· 2 小時前AI 評分40
AI 導讀
令人印象深刻的 NanoGPT 新歷史紀錄,39.9 秒 (-27.7 秒) 來自
@DevenPzak, 徹底擊敗先前 67.6 秒的紀錄!這個紀錄為 NanoGPT 引入了新的思考範式:不再最佳化矩陣乘法或加入更具表現力的操作,而是以單個 flop 為單位進行最佳化,並
正文
impressive New historic NanoGPT record at 39.9s (-27.7s) from
[@DevenPzak](https://x.com/DevenPzak)
, obliterating the prior record of 67.6s! This record introduces a new paradigm of thinking to NanoGPT: instead of optimizing matmuls or adding more expressive operations, optimize at the individual flop level with
來源:Thom Wolf (@Thom_Wolf) · x.com