Warning message

Submissions for this form are closed.

Tensorflow and Deep Learning, 2019-05-(08-09)

Tensorflow and Deep Learning


This two day course is made up of two parts. The first part will have an introduction to Tensorflow, give some basic knowledge about Generative Adversarial Networks (GAN), and teach how to use GAN advanced deep learning in practice for some fancy and useful applications. The other part of this course will introduce some of the recent advanced deep learning models (e.g., Transformer, ELMo, BERT) in NLP (natural language processing).

The course consists of lectures and hands-on exercises. Google CoLab (or your own Jupyter Lab) environment will be the working environment in the course. HPC2N's cluster Kebnekaise will be used as a backup resource in case the first two options does not work.


Day 1: Tensorflow, and Deep learning models in NLP-1

  • 9:00-10:00  Lecture: Introduction to Tensorflow.
  • 10:00-10:30 Exercise: Play with Tensorflow and TensorBoard on basic.
  • ~10:30 Coffee break
  • 11:00-12:00 Exercise: Play with Tensorflow for a regression application on a real dataset.
  • 12:00-13:00 Lunch
  • 13:00-14:00 Lecture: Introduction of Deep Learning on Natural Language Processing (NLP).
  • 14:00-14:30 Exercise: Running data representation in basic.
  • ~14:30 Coffee break
  • 15:00-16:00 Exercise: Play with the deep learning model in practice.

Day 2: Deep learning models in NLP-2 and Generative Adversarial Networks (GAN)

  • 9:00-10:00 Lecture: Introduction of different state-of-the-art deep learning models on NLP.
  • ~10:15 Coffee break
  • 10:30-11:45 Exercise: Play with the deep learning models in practice in practice.
  • 11:45-12:00 Lecture: Wrap-up of Deep Learning Models in NLP.
  • 12:00-13:00 Lunch
  • 13:00-14:00 Lecture: Introduction of Generative Adversarial Networks (GAN).
  • 14:00-15:15 Exercise: Play with GAN basic on real dataset.
  • ~15:15 Coffee break
  • 15:45-16:15 Wrap-up: Get to know the training challenges in training GAN, as well as the GAN advanced models.


Please bring a laptop. Please make sure to follow the installation requirements in below link to set up the working environment before the course.


If there is any trouble, please create an issue within the repo.

Participation is free. Lunch and coffee/tea will be provided.

Note: Participants are kindly asked to limit the use of fragrances due to perfume intolerance issues.  Thank you.

Prerequisites: Basic knowledge of a Linux/Unix environment will be assumed. The participants are also assumed to have attended the following HPC2N previous course: "Introduction to Deep Learning" or have experience with the contents of that course (especially Day1).

Slides from the course "Introduction to Deep Learning", made by Markus Koskela, CSC, can be found here:

  • Day 1 (1, 2, 3)
  • Day 2 (4, 5).

Instructors: Lili Jiang, Xuan-Son Vu, Michele Persiani.

Time and date: 8-9 May, 2019. 9:00-16:30 both days.

Location: N280 (Naturvetarhuset), Umeå University. Getting to Umeå University/HPC2N: info here.

Deadline for registration: 1 May, 2019. The number of seats are limited, and this means the registration may be closed earlier, in case the space fills up. REGISTRATION CLOSED! All seats taken. We will have a waiting list in case someone cancels. Please send an email to bbrydsoe@hpc2n.umu.se if you want to be added to the waiting list.

Please register by filling in the below form. Fields marked with a * are mandatory.

Updated: 2023-09-11, 11:36