At Leipzig Python User Group, Meetup / Homepage, 2018-02-13 19:00 at Basislager.

We looked briefly at the history of deep learning and one modern deep learning framework: PyTorch.

• Slides - source

Notebooks we worked through:

1. Benchmark: Python vs NumPy vs PyTorch
2. Basic PyTorch: The tensor class, numpy-torch interop
3. Automatic differentiation: A brief excursion into autograd, a library that can automatically differentiate native Python and Numpy code.
4. Basic Multilayer Perceptron: A basic multilayer perceptron learning example with random data.
5. CIFAR10: A convolutional neural net for recognizing objects in small images.