註冊並分享邀請連結,可獲得影片播放與邀請獎勵。

Emily
@IamEmily2050
Any sufficiently advanced technology is indistinguishable from magic. Arthur C. Clarke.
加入 October 2014
1.1K 正在關注    49.2K 粉絲
Nano Banana Pro V2 is that you 👁️ In this Gemini Notebook video overview, we'll look at the new research from the Google DeepMind team on Elastic Looped Transformers (ELT). This innovative visual generation framework replaces conventional deep transformer stacks with iterative, weight shared recurrent blocks. To address the typical opacity of intermediate layers, the authors propose an Intra Loop Self Distillation (ILSD) training strategy that effectively guides early student loops using a maximum-loop teacher configuration. This architecture decouples physical parameter size from computational depth, enabling a single trained model to perform Any Time inference with dynamic trade offs between generation quality and speed. Consequently, ELT achieves state of the art parameter efficiency and faster training convergence while significantly reducing memory transfer bottlenecks in both image and video synthesis tasks.
顯示更多