00:32:22 Anon. Sanger’s Rule: no 00:32:24 Anon. Kirchhoff: no 00:32:26 Anon. Gaussian: no 00:41:49 Anon. GAN: Shouldn't the first output after translation be ? 00:42:37 Akshat Gupta (TA): Its the first input to the decoder module 00:42:51 Anon. GAN: Got it Thanks! 00:44:08 Anon. Activation Function: randomly select based on the distribution? 00:44:36 Anon. CTC: np.random 00:44:36 Anon. GAN: uniform distribution 00:49:43 Anon. Gaussian: no 00:51:31 Anon. Linus: No 00:51:32 Anon. Activation Function: no 00:51:33 Anon. Sum: no 00:51:37 Anon. GAN: nope 00:51:39 Anon. Graph: qed 00:51:41 Anon. Activation Function: nose 00:51:54 Anon. GAN: he nose dived 00:52:19 Anon. Backprop: the problem is nosedive is one word 00:52:34 Anon. GAN: yeah.. just a random suggestion.:) 00:52:44 Anon. Center Loss: great suggestion 00:52:49 Reshmi Ghosh (TA): Good one 00:53:07 Anon. Backprop: Yeah it was still a good guess 00:53:24 Reshmi Ghosh (TA): Yuss clear all worked 00:53:27 Reshmi Ghosh (TA): Shoot! 00:53:38 Reshmi Ghosh (TA): Ignore plis, message was for akshat 01:05:51 Anon. Jacobian: add a length constraint on the output 01:05:52 Anon. Jacobian: ? 01:07:25 Reshmi Ghosh (TA): Sorry what is add a length constraint? 01:07:39 Reshmi Ghosh (TA): We clearly missed this 01:38:26 Reshmi Ghosh (TA): I am proud of everyone paying “attention” in class today because I cannot stop myself from checking the poll results. 01:39:26 Anon. Gh0stR1d3r: we haven’t had polls yet?? 01:39:28 Anon. Gh0stR1d3r: oh wait 01:39:36 Anon. Momentum: Multithreading heavily lol 01:39:49 Reshmi Ghosh (TA): XD 01:40:00 Anon. Gh0stR1d3r: doom scrolling is a hungry process y’all don’t do it 01:40:06 Reshmi Ghosh (TA): No polls because main polls at stake 01:40:18 Reshmi Ghosh (TA): No one had the time to create it yesterday 01:42:56 Anon. Caffe: Dooming +1