Introduction to Deep Learning
ECTS: | 6 |
Semester hours: | 4 |
Exam mode: | Written |
Language: | En |
Links: | CIT, Courses |
Exams: |
2023WS
Endterm
2023SS Endterm 2022SS Endterm 2021WS Endterm 2021SS Endterm 2020WS Endterm 2020SS Endterm 2019WS Endterm 2018WS Endterm 2018SS Endterm |
Reviews
Has anyone experience how much workload the homework exercises are for this course?
- I have. Usually, first homework is easy, but second and third are dramatically difficult.
I do remember professor laura mentioned that this semester onwards she will remove the project in the end and make it more homework intensive.
Last semester there were only 3 homeworks and 1 project work.
Just make sure to start early, work on python, browse code of pytorch to understand how Resnet, Alexnet and other common deep learning models are built.
Thanks for the info. Do you know by chance what’s the exam gonna be like?
- Yeah. Its tricky questions, things that you wont get to practice, scoring above 2.7 is really easy, but going higher will be difficult.
A lot of tiny numericals, such as computing output of size for a given convolution filter, upconvolution filter, backpropogation. MCQ with right answers(if unticked is correct, and you leave it unticked, you get marks there too), etc.
Basically, for anything you study, there will be a numerical or conceptual question.
Most focus is on CNN and RNN,LSTM.
I would recommend studying from CS231n Stanford CNN course.
And attend the classes, many dont do that, and you can bet a few marks loss there. Cause sometimes prof will say something which is not in the slides.