All entries for March 2020

Screen casting while working from home

Writing about web page https://www.howtogeek.com/446706/how-to-create-a-screencast-on-linux/

As I am sitting at home due to the Covid-19 pandemic and the closure of "outside life", I find myself having to give feedback to PhD students who are not around. Normally, I would do this with them sitting next to me. Now, I am trying to do it using a "Screencast", i.e. a video that shows my screen and records my voice, perhaps even my video, when I go through their work and tell them what I think about it and how it could be improved. I have found that under Linux, there is a varirty of useful and free such screencast software. After a bit of googling, I found

https://www.howtogeek.com/446706/how-to-create-a-screencast-on-linux/

and at the moment, I am using the "recordmyDesktop" tool which one can install via the ususal Linux software channels. After a few tries, I think this is indeed a great tool. I'll let you know what the students think ...

Rudo Roemer : 31 Mar 2020 12:47 |  Tags: Casting Linux Screen |  Comments (0) |  Close comments |  Report a problem

list of submitted projects and information about late submission penalties

Follow-up to Submission date and instructions for your project work from Rudo's blog

Dear all,

thus far I have received submissions from (10/02/30 10:00)

1. Jettawat Khantaveramongkol
2. Juan Roman Roche
3. Hendrik Schlueter
4. Alexios Malaquias
5. DanielPepin Fowan
6. Burak Civitcioglu
7. Roman Perrier (one day late)
8. Jeremy Cadran (two days late)
   

Please note that I intend to apply late submission penalties, following roughly similar Warwick penalties which penalize a late submission with a 5% reduction in final points for every day late. A late day starts in the first minute after the dealine and the time of submission follows from the time of the last upload.

Regards,

-Rudo

Rudo Roemer : 05 Mar 2020 13:16 |  Tags: Fastai Keras Machine-Learning Tensorflow |  Comments (0) |  Close comments |  Report a problem

PS: don't worry too much about maximizing epochs

Follow-up to Submission date and instructions for your project work from Rudo's blog

Just to add: do not worry too much about running your code to the maximum of epochs. It is clear to us that most people do not have access to state of the art graphics cards or hundreds of processors. What is more important is to see that you came up with a good idea where machine learning can be used, executed the right commands to get some training and analysis done, and found the right data to train on. In fact, we migth even try to run some of your code for more epochs to see how results can get better!

Rudo Roemer : 02 Mar 2020 10:26 |  Tags: Fastai Keras Machine-Learning Tensorflow |  Comments (0) |  Close comments |  Report a problem

Submission date and instructions for your project work

Follow-up to teaching schedule, now ready from Rudo's blog

Dear all,

you should now all be close to finishing your project work for the machine learning course. I hope it has all gone well and that you had some fun exploring machine learning. The official date for submission is

Friday, March 6 at 17:00.

For submission, please remember that we want you to submit a running machine learning code using a jupyter notebook with comments that explain step by step what you did and why. Please also supply the data so that Djena and I can run the code ourselves. The easiest way to send us the submission is via

https://files.warwick.ac.uk/rroemer/sendto

where you can upload both jupyter notebook and data (preferably as a single .zip file). If your data files are too big for upload there (I have 4.5GB in total on that upload drive), please send a link at which we can then download the data. Please remember that it is your responsibility to make sure data and notebook work "out of the box", i.e. without us having to change anything before they work, including directories.

We aim to have feedback for you by the end of March and will communicate this by email to each of you (so please also let me have your preferred email address in your jupyter notebook).

Cheers!

Rudo Roemer : 02 Mar 2020 09:08 |  Tags: Fastai Keras Machine-Learning Tensorflow |  Comments (0) |  Close comments |  Report a problem