interactiveaudiolab / course-deep-learning Goto Github PK

There's a lot of great info in here. I did some mild (very mild) reorganization. The main thing I see is that it might cover too many topics in one notebook and we may want to split this up into 2 notebooks. I'm thinking about how we want to do this....I can imagine putting everything after the first test/train loop (section 4) into a 2nd notebook that contains GPU, logging, and data augmentation.

Also that main training loop gets repeated multiple times in different cells. I feel like there should be some way of declaring it a single time and then calling it repeatedly...

Another issue is that the Tensorboard logger didn't display anything for me. Maybe I did something wrong. Not sure.

Also, I feel like we introduce an optimizer (SGD) but don't really talk about it and other elements (e.g loss functions) were explicitly talked about.

OK. Those were my reactions. Now we have stuff to discuss!

Cloning into 'course-deep-learning'...
remote: Support for password authentication was removed on August 13, 2021. Please use a personal access token instead.
remote: Please see https://github.blog/2020-12-15-token-authentication-requirements-for-git-operations/ for more information.
fatal: Authentication failed for 'https://USERNAME:[email protected]/interactiveaudiolab/course-deep-learning.git/'
[Errno 2] No such file or directory: 'course-deep-learning/'
/content
ERROR: Could not open requirements file: [Errno 2] No such file or directory: 'requirements.txt'

Hmm...so when a point is near to the plotted decision surface...but on the wrong side of the line... the perceptron doesn't update the decision surface. Is there a problem with the decision surface plotting or is there a problem with the perceptron code? Not sure. Have a look at this.

I couldn't download audio mnist. See the attached image for the details

.