Worth
《AN IMAGE IS WORTH 16X16 WORDS: TRANSFORMERS FOR IMAGE RECOGNITION AT SCALE》阅读笔记
论文标题 《AN IMAGE IS WORTH 16X16 WORDS: TRANSFORMERS FOR IMAGE RECOGNITION AT SCALE》 谷歌论文起名越来越写意了，“一幅图像值16X16个单词” 是什么玩意儿。 AT SCALE：说明适合大规模的图片识别，也许小规模的不好使 ......
Firms sign investment deals worth $8.6b at Invest Beijing Global Summit
Beijing will continue to be an attractive investment destination for foreign investors, as the city is at the forefront of technological innovation, h ......
Sep 2022-Prioritized Training on Points that are Learnable, Worth Learning, and Not Yet Learnt
提出了Reducible Holdout Loss Selection (RHOLOSS)，一种简单但有原则的技术，近似地选择那些最能减少模型泛化损失的点进行训练 ......
An Image Is Worth 16x16 Words: Transformers For Image Recognition At Scale
模型如下图所示: 将H×W×C的图像reshape成了N×(P2×C),其中(H,W)是图像的原始分辨率,C是通道数,(P,P)是每个图像块的分辨率,N=H×W/P2为图像块的数量,将一个图像块使用可学习的线性层映射到维度为D的隐藏向量,如式(1)所示,线性映射的输出称为patch embeddin ......