00:01:20.475,00:01:23.475
umberto valleriani: Yes

00:01:22.894,00:01:25.894
Marzia Riso: Yes

00:02:39.094,00:02:42.094
umberto valleriani: I didn't

00:02:58.474,00:03:01.474
donato crisostomi: I did 

00:03:05.886,00:03:08.886
francesco veneziano: yes

00:03:06.942,00:03:09.942
valerio mannucci: yes

00:03:10.913,00:03:13.913
riccardo taiello: yes

00:03:21.605,00:03:24.605
daniele baieri: Me too

00:03:26.690,00:03:29.690
Simone Antonelli: yes

00:03:34.086,00:03:37.086
andrea caciolai: yes

00:03:38.636,00:03:41.636
mattia capparella: I did send th3 solution to prof Rodolà

00:04:07.531,00:04:10.531
donato crisostomi: yes give me one second to get colab working

00:05:31.699,00:05:34.699
mattia capparella: Ok

00:05:46.621,00:05:49.621
donato crisostomi: yes, almost done

00:10:36.910,00:10:39.910
shadi andishmand: I don’t have any voicevoice

00:10:50.831,00:10:53.831
Emanuele Rodola': in the sense that you can't hear us?

00:11:11.388,00:11:14.388
umberto valleriani: Einstein sum notation isn't really clear for me

00:11:24.671,00:11:27.671
shadi andishmand: Yes i was hearing but now anything

00:11:42.220,00:11:45.220
andrea caciolai: umberto try this link https://stackoverflow.com/questions/26089893/understanding-numpys-einsum

00:11:45.454,00:11:48.454
umberto valleriani: I think i understand the basics but don't understand the explicit form

00:11:55.977,00:11:58.977
andrea caciolai: it helped me out a lot understanding

00:12:05.483,00:12:08.483
umberto valleriani: Thanks

00:13:01.466,00:13:04.466
Emanuele Rodola': shadi, try to logout and login again, and try to close all the applications that use sound / mic

00:16:28.104,00:16:31.104
umberto valleriani: Ok thanks a lot

00:20:08.987,00:20:11.987
leonardo torcellini: how do you tell which is the max size your gpu can fit?

00:21:26.622,00:21:29.622
Emanuele Giona: i usually try to instantiate the model with a given batch size, if it works i try to increase it until i get a memory error during allocation

00:25:06.488,00:25:09.488
leonardo torcellini: ok thanks

00:26:33.672,00:26:36.672
denis stanev: Yes, we can see it.

00:31:01.224,00:31:04.224
Luca Moschella: https://discuss.pytorch.org/t/what-is-adaptiveavgpool2d/26897/2?u=ptrblck

00:31:46.799,00:31:49.799
Marzia Riso: Got it, thanks!

00:31:47.476,00:31:50.476
shadi andishmand: no its clear

00:31:50.590,00:31:53.590
shadi andishmand: thanks

00:45:26.974,00:45:29.974
shadi andishmand: professor could you explain a little about reproducibility when we run same data on same algorithm but we have differnt result sometimes! and when we should do it or before which algoritms? for ex in kfold or random forest e have a parameter random_state. so does it effect or not?

00:45:56.611,00:45:59.611
Emanuele Rodola': ok noted

00:46:32.267,00:46:35.267
francesco veneziano: i see your screen too small

00:46:43.680,00:46:46.680
francesco veneziano: it's ok

00:46:46.871,00:46:49.871
Emanuele Rodola': francesco i see it perfectly

00:46:48.266,00:46:51.266
Emanuele Rodola': ok

00:47:03.112,00:47:06.112
valentina de paolis: me too

00:49:31.097,00:49:34.097
francesco veneziano: yes thanks

00:53:14.185,00:53:17.185
shadi andishmand: thanks a lot .so when intialise seed the weights update each time in same huristic way? i mean backpropagation part? 

00:53:17.615,00:53:20.615
denis stanev: How big of a performance impact does the deterministic mode have?

00:53:19.781,00:53:22.781
shadi andishmand: yes

00:53:25.290,00:53:28.290
Valerio Neri: Is it a good practice to set the random seeds manually on a project? Should we always set them?

00:54:48.694,00:54:51.694
shadi andishmand: thanks a lot professor

00:55:29.285,00:55:32.285
shadi andishmand: yes it is slowing working even with gpu

00:55:35.705,00:55:38.705
valerio mannucci: shoulden't the error for validation set  increase a lot when the degree became big?(about last graph)

00:56:01.917,00:56:04.917
valerio mannucci: i can't see in last graph

00:56:32.864,00:56:35.864
valerio mannucci: ok thanks

00:59:02.378,00:59:05.378
francesco veneziano: yes, but the problem pointed out in jupyter, not in colab

01:00:06.972,01:00:09.972
francesco veneziano: ok thanks

01:00:14.666,01:00:17.666
umberto valleriani: Just a casual question: in the last lesson (about CNNs) you mentioned we're going to see some applications on graphs. Could you anticipate a bit on what are they?

01:01:44.613,01:01:47.613
umberto valleriani: Sounds cool, thanks

01:05:33.969,01:05:36.969
shadi andishmand: some libraries like plotly have some problem in pycharm. i use scientific mode of pycharm and can plot by matplotlib but i imported plotly and it was successful . i run the project without any error but i cant see the plot that i can see in jupyter or google colab. i wanted to ask its about pycharm limitation in ploting with some library?

01:06:46.889,01:06:49.889
shadi andishmand: thanks a lot 

01:08:06.575,01:08:09.575
Andrea Rosati: lol

01:10:47.687,01:10:50.687
valerio mannucci: could you explain the utility of detach() in a learning model

01:10:47.736,01:10:50.736
dario aragona: when saving colab notebooks I often get an "automatic save pending for X minutes" and it stucks

01:10:52.513,01:10:55.513
Luca Moschella: Shadi I re-read your question, if you have problems even to only show the plot in the browser tab, be sure to call the `.show()` on the fig obj

01:11:02.968,01:11:05.968
dario aragona: does anybody else get the same error?

01:13:06.378,01:13:09.378
valerio mannucci: thanks

01:13:08.976,01:13:11.976
shadi andishmand: i did that luca . i have plot for matplotlib but no plot for pltly even with no error. i should say i use ubunto.  if i dont find any good way to ploting maybe i use google colab . thanks

01:14:49.694,01:14:52.694
dario aragona: it's strange because my connection works, and also if I manual save it  gets stuck

01:14:53.852,01:14:56.852
dario aragona: but I'll google it

01:14:56.859,01:14:59.859
dario aragona: thanks!

01:15:02.436,01:15:05.436
Emanuele Rodola': also try changing browser

01:15:04.112,01:15:07.112
dario aragona: yes I use Chrome

01:15:04.610,01:15:07.610
Emanuele Rodola': right

01:15:17.946,01:15:20.946
Emanuele Rodola': try disabling AdBlock

01:15:38.968,01:15:41.968
dario aragona: Yes, I download the notebook and upload it on my drive

01:15:45.621,01:15:48.621
dario aragona: Ok I'll check

01:17:09.450,01:17:12.450
shadi andishmand: thanks a lot professor . happy easter.

01:17:19.055,01:17:22.055
Emanuele Giona: goodbye prof

01:17:19.195,01:17:22.195
francesco veneziano: it's ok, thanks professors. Happy easter

01:17:19.876,01:17:22.876
dario aragona: Thank you, happy holidays

01:17:22.463,01:17:25.463
mohammad kamel: Thank you. bye

01:17:23.977,01:17:26.977
donato crisostomi: thanks, goodbye and happy holidays

01:17:25.313,01:17:28.313
umberto valleriani: Bye thanks

01:17:26.215,01:17:29.215
Simone Antonelli: Goobye

01:17:26.450,01:17:29.450
andrea caciolai: thanks goodbye

01:17:27.692,01:17:30.692
davide sforza: Bye!

01:17:27.739,01:17:30.739
daniele baieri: Okay, thank you very much. Bye

01:17:28.766,01:17:31.766
Antonio Musolino: Goodbye

01:17:28.815,01:17:31.815
riccardo taiello: thanks bye

01:17:28.967,01:17:31.967
Valerio Neri: Goodbye

01:17:29.572,01:17:32.572
denis stanev: Goodbye.

01:17:30.052,01:17:33.052
seid koudia: thank you bye 

01:17:31.672,01:17:34.672
leonardo torcellini: goodbye

01:17:36.207,01:17:39.207
stefano esposito: thanks and goodbyw

01:17:37.095,01:17:40.095
Marzia Riso: Thank you so much! Have a nice day

01:17:37.141,01:17:40.141
luca podo: thanks bye

01:17:38.408,01:17:41.408
mattia capparella: Bye bye! Happy holidays!

01:17:40.790,01:17:43.790
shadi andishmand: goodbye luca and antonio . happy easter too

01:17:42.679,01:17:45.679
valentina de paolis: good bye

01:17:43.027,01:17:46.027
shadi andishmand: bye