Important links

Dataset:

https://www.kaggle.com/datasets/edenbd/150k-lyrics-labeled-with-spotify-valence

Project plan:

https://docs.google.com/document/d/1bslj6MDCvOimTr_xIhf93SIW2_bE3bcZ/edit

Project timetable:

https://docs.google.com/spreadsheets/d/1yx6rCDp3FSVsgFSkH1lefZThQxoNE3jY/edit?gid=765237529#gid=765237529

Colabotary

https://colab.research.google.com/drive/1-1C77gQeQutaZmnxjXAn63HdtHz9Bxnf?usp=sharing

Notebook with VM how 2

Running a jupyter notebook

1. Log into hpc with your username, you can use PuTTY or (open)ssh.
2. you may change the resources you request by changing the lines that are starting with #SBATCH. Currently, it runs for 1h has 4 cpus and 8GiB of ram.
3. copy the ./jupyter_root.sh to the vm
4. run sbatch -A ealloc_mtat-03--1 jupyter_root.sh
5. run cat jupyterlogs/jupyter_root.log (you may have to wait a bit, till it runs)
   • you need the following lines:
     • "ssh -N -L 8044:172.16.10.124:8044 hpc_sadneutrino@falcon4.hpc.ut.ee" (ports and username is different)
     • "http://127.0.0.1:8044/tree?token=f72ac1855426c53208dbfa8dbb08d9dfa6f01f64b4a88321" the last https link with 127.0.0.1
     • "localhost:8044"
6. copy and paste the ssh command to a terminal (vscode)
7. click on the python<x.y> and if it asks for url paste the http url that you copied
8. open the "localhost:port" in the browser, and log in with: part of the url after = here it is "f72ac1855426c53208dbfa8dbb08d9dfa6f01f64b4a88321"
9. upload the requirements.txt into the folder you are in

Killing the jupyter notebook

1. run squeue -u $(whoami)
2. copy the jobid
3. run scancel <jobid>