Lesson 1 Discussion βœ…

(Arul Bharathi) #551

Amazing lecture @jeremy. Looking forward to learn about how to create our own dataset from google images. Thanks!

0 Likes

(Bart Fish) #552

gotta have label, this is Supervised training. Unsupervised training is possible, but that’s in Part 2.

1 Like

(jaideep v) #553

if we do nn.parallel what different would it do

0 Likes

(Matej) #554

Hi Jeremy, can I rewatch the todays lecture? Its been during the night here in EU, and I was not as awake as I wanted to be …

1 Like

(Bharat Kunwar) #555

runs the model on multiple-gpus

0 Likes

(Kartik Sharma) #556

Language Model Zoo 🦍 this one ?

search for ULMFit

0 Likes

(Marvin Petersen) #557

Learning rate is too high. So parameter updates are inadequate. Neatly explained here https://towardsdatascience.com/estimating-optimal-learning-rate-for-a-deep-neural-network-ce32f2556ce0

0 Likes

(Pierre Ouannes) #558

No problem :slight_smile: So is it something like what @KevinB said, that this number is actually the prediction percentage for the incorrect class (and not the correct one like Jeremy said) ? But if it’s not I really don’t understand this point…

0 Likes

#559

Great session, also a huge thanks to the people answering questions live on here!

1 Like

(yogesh gurjar) #560

is this resolved anytime soon or can i do some custom changes so it work for me on my colab fastai functions

0 Likes

(Fadhli Ismail) #561

If I’m not mistaken, the video will be available right away after the lecture

0 Likes

(Hooin Kyoma) #562

correct

0 Likes

(Stas Bekman) #563

No, idea, I was just linking to where it’s discussed, please continue over there

0 Likes

(Daryl Wilson) #564

Thanks Jeremy and Rachel, excellent lecture.

0 Likes

#565

the learn.lr_find() finds the optimal rate for which layers?

2 Likes

(George Zhang) #566

Pay attention to the fact that ResNet-50 takes less around half of the operations to achieve equally competitive results. So though I agree that right now Inception-v4 is the best if you want to be very accurate, but ResNet wins in terms of efficiency.

1 Like

(Leo Yu Ho, Lo) #567

Any threads that have instructions for using multiple GPU with fastai library?

I searched, but not the one that with instructions.

Thank you in advance!

0 Likes

(Pawan) #568

Thanks a lot!

0 Likes

(Bart Fish) #569

search for nn.DataParallel

1 Like

(Bharadwaj Krishnan) #570

Just want to make sure. the stream link (https://www.youtube.com/watch?v=7hX8yKCX6xM) will still be re-watchable through the length of the course(till dec) right? or would it be taken down?

2 Likes