The first version of HW2 contained one erroneous test (the last one). The reason for the error was that in the reference solution we used int16 for POS transition counts which overflew for some of the transitions. We have corrected this mistake and added public tests that raise an error if the count matrices and the probability matricies have negative elements. These changes only affect Exercise 4 and 5. The instructions did not change, your old solution should work fine (if it is correct).
If you click on the invitation link after April 2, Monday, you will automatically clone the corrected version.
If you created your own homework repository before April 3, you have to
manually pull the changes. If you already have changes in the file (you started
to work on the homework), the easiest way is to copy your existing solution to
a separate file, and reset homework2.ipynb
before pulling (you do not want to
resolve a merge conflict between Jupyter Notebooks).
Copying and resetting the file in command line interfaces (Linux Terminal, MacOS Terminal and Powershell):
cd /path/to/homework2-johnsmith
cp homework2.ipynb homework2-old.ipynb
git checkout -- homework2.ipynb
Now you can add the common repository as a new remote, fetch and merge the new version:
git remote add python-nlp https://github.com/python-nlp/homework2.git
git fetch python-nlp
git merge python-nlp/master
You can copy and paste your old solution into the new notebook.
If you have done everything correctly, the last assert statement should be this:
assert tags == ['DT', 'NN', 'VBZ', '.']
This page describes general information regarding the mandatory homeworks for this course.
The actual homeworks are published in the homeworks/
directory.
Please read the instructions and README.md
(if available) files for each homework carefully.
The homeworks are individual assignments. Collaboration is not prohibited but you are expected to submit your own solution. You may be required to orally present your solutions.
Submissions are managed through Github Classroom. You will receive an invite link via Neptun.
- Click on the invite link and login to GitHub. Create an account if necessary.
- Grant access to GitHub Classroom. It will only use your public information and your homework repository's information.
- Find your name in the course roster and click on it. If you can't find your name in the list, please email the instructor.
Now Github Classroom creates your own private repository for this homework with the starter code (the homework notebook) in it. This may take a few minutes.
If you're unfamilar with git, many great tutorial are available online such as the Git Book here (Chapter 1 and 2 should be more than enough for this course).
The last commit in this repository before the deadline will be your submission for this homework.
Please be aware that we might delete these repositories after the end of the semester.
One homework assignment can be (re-)submitted by the end of the repeat period in accordance with the Code of Studies and Exams.
Each homework is graded from 1 to 5. You are required to achieve the grade 2 for each homework in order to take the final exam. The scoring system is released along with each homework.
All deadlines are Sunday 23:59 PM CET.
Homework | Release Week | Deadline Week | Person in charge | |
---|---|---|---|---|
1 | 3 | 6 | Judit Acs | [email protected] |
2 | 7 | 10 | Judit Acs | [email protected] |
3 | 10 | 13 | David Nemeskey | [email protected] |
If you have any questions or general feedback regarding the homeworks, please file a GitHub issue with the homework
tag.