回到意大利之後，安迪把「蛋炒飯」作為一項重要技能引進介紹給身邊的朋友，認為「把早餐的面包替換成這個能對自己稍微好點」。
Fixed/sinusoidal positional encodings are not counted (following the original Transformer paper convention)
，详情可参考91视频
Cross-layer sharing, rank-1 projections, sparse gate, low-rank head, frozen scaling params
Streamer said ‘deal no longer financially attractive’ at price required to match offer by David Ellison’s firm