Abstract:

In this talk, we will present the basics of PyTorch, including Tensors, automatic differentiation with autograd and the neural networks package. We will also present an overview of torchvision, a library that extends PyTorch with features specific to Computer Vision, such as pre-trained models for several tasks, built-in datasets, data transformation and video reading.

Short bio:

Francisco Massa is a Research Engineer at Facebook AI in Paris, France. Prior to joining Facebook, he was a Research Engineer with the Magic Pony team at Twitter (London). Francisco received his PhD in Computer Vision in 2017 from Université Paris-Est, France, under the supervision of Mathieu Aubry and Renaud Marlet, and his B.Sc and M.Sc. in 2012 from École Polytechnique, France. At Facebook, Francisco works on the PyTorch team, with a focus on Computer Vision.

Deep Learning & PyTorch at UPC School:

If interested, check our PyTorch labs developed by Santiago Pascual, Daniel Fojo and Albert Pumarola for our postgraduate program in Artificial Intelligence with Deep Learning at UPC School.