nyu-dl / nlp_dl_lecture_note Goto Github PK

Contributors

Stargazers

Forkers

multipath ivanhe longyuewangdcu icemansina jgutman adrianhust cherrypotter niangaotuantuan peratham jermainewang xpapag ilovejs tigerneil amoliu fooway nonva little1tow eshijia jeffzhengye shannonyu tfwu programmer-util asn264 vikingmew joebradly jpcolomer ankitkv shashankg7 sigmaquan ompanda worldpress codeaudit manparvesh yochju jiangnanhugo sleepinyourhat zijialewislu fducau ratishsp xc918 vyraun zshwuhan recluze tonydeep datavizweb zxsted rajusamantapudi keunwoochoi johndpope rubythonode charlie13 hephaex mdbishop hades210 vikramkaku jlee12393 aodongli liyc7711 vkhokhla christinahyc yaningfords mridul timwee nipandha ychernushenko mohanarunachalam galacticmclazy shubhamrohilla05 fengyinyang colinsongf barak zoema guozhiqi14 521314 hal2001 mkmojo yucoian ls4408 harshlal028 uralik johnyjyu fguan322 codingdoctor gilcreativity afcarl yjo2lee kehang jungsuahn cl2534 evaaaaaaa nicoledjy tatenda-ndambakuwa zihao-roy yx1201 leesehoon xiaoyizy jbrry skoundin abiraja2004 urielll

\frac{\partial h_{t}}{\partial \underline{h}_{t}}

DU = Q (\diag(D) \odot S) Q^{-1}

\diag(S^{l-t'}) \odot (Q Q^{-1} W\phi(x_{t'}))

S^{l-t'} Q Q^{-1} W\phi(x_{t'})

Q S^{l-t'} Q^{-1} W\phi(x_{t'})