<div style="line-height:1.7;color:#000000;font-size:14px;font-family:Arial"><div style="margin:0;">Hi Adapters,</div><div style="margin:0;"><br></div><div style="margin:0;">As large-scale pre-trained language models are ubiquitously used in various NLP tasks, the tremendous computational cost incurred by such models hinders their practicality in resource-constrained and time-sensitive scenarios. Therefore, a plethora of acceleration methods emerged by tackling different aspects of model redundancy. In this talk, I will introduce two categories of model acceleration algorithms: sample-adaptive early-exit and layer-adaptive length reduction. The former is built upon the idea that different samples have different difficulty, while the latter focus on in-sample redundancy among tokens.</div><div style="margin:0;"><br></div><div style="margin:0;"><div dir="auto" style="font-family: Helvetica, "Microsoft Yahei", verdana; word-break: break-word !important;"><b style="word-break: break-word !important;">Time:</b> 2022/04/06 16:00-18:00</div><div dir="auto" style="font-family: Helvetica, "Microsoft Yahei", verdana; word-break: break-word !important;"><br style="word-break: break-word !important;"></div><div dir="auto" style="font-family: Helvetica, "Microsoft Yahei", verdana; word-break: break-word !important;"><b style="word-break: break-word !important;">Venus: </b></div><div dir="auto" style="font-family: Helvetica, "Microsoft Yahei", verdana; word-break: break-word !important;">Tencent Meeting Link: <a href="https://meeting.tencent.com/dm/qDcCYYV2YUIj" style="color: rgb(6, 73, 119); word-break: break-word !important;">https://meeting.tencent.com/dm/qDcCYYV2YUIj</a></div><div dir="auto" style="font-family: Helvetica, "Microsoft Yahei", verdana; word-break: break-word !important;">Tencent Meeting Number£º202-706-320</div><div dir="auto" style="font-family: Helvetica, "Microsoft Yahei", verdana; word-break: break-word !important;"><br style="word-break: break-word !important;"></div><div dir="auto" style="font-family: Helvetica, "Microsoft Yahei", verdana; word-break: break-word !important;">Hope you will find it useful and interesting!</div><div dir="auto" style="font-family: Helvetica, "Microsoft Yahei", verdana; word-break: break-word !important;"><br style="word-break: break-word !important;"></div><div dir="auto" style="font-family: Helvetica, "Microsoft Yahei", verdana; word-break: break-word !important;">Best regards,</div><div dir="auto" style="font-family: Helvetica, "Microsoft Yahei", verdana; word-break: break-word !important;">Roy</div></div></div><br><br><span title="neteasefooter"><p> </p></span>