Hands-On Practice for your Deep Learning Skills - [IPython Notebooks]

cristi (70)in #deep-learning • 7 years ago

Kyubyong Park posted on his github a series of exercises for those who want to polish their deep learning skills with Pytorch.

Looking into his repositories I realize that he has many similar folders; he's got exercises to improve your TensorFlow skills, and he also has some exercises on Numpy. So, you should check them out if this is in your interest.

As for those about Pytorch, he has divided them into two chapter (I suspect this is ongoing work and he may add some more). According to Park:

"Pytorch is one of the most popular deep learning libraries as of 2017. One possible way of familiarizing yourself with it, I think, is to practice with simple quizzes. That's where this project comes in." [source]

The first chapter is on tensors and you'll practice on:

how to create and transform tensors
what is random sampling
maths and statistics
and more.

The second chapter is on neural networks and is more inclusive:

linear, convolution, recurrent, pooling, dropout and other layers
activation functions
loss, distance, initialization, optimization
and some utilities.

The exercises and the solutions are currently only available for the first chapter, while the second is still 'work-in-progress'. So, what are you waiting for? Go ahead and practice:

Hands-On Practice for your Deep Learning Skills - [IPython Notebooks]