(Slow) Virtual Study Group: Karachi (Pakistan)

(This seems like the best time zone group where I can put myself out there / onto the forums. Apologies if it is for India only; I will try to post elsewhere if so.)

I am making a career transition from qualitative research to something more quantitative. I’ve been studying software engineering online for the past year or so (mainly Ruby, but some Python and JavaScript) but always wanted to dive into deep learning and machine learning / data science work.

(Some basic background in case of interest: I’m 36 years old, Dutch / British background. Living in Pakistan at the moment but I used to travel a lot pre-COVID. My background is in history / qualitative research work, in which I got my PhD a few years back.)

I’m interested in seeing if there is anyone in or around this time zone — I’m based in Karachi, Pakistan, for the moment — who would be interested in chatting together or working together through the fast.ai syllabus / course? Ideally a first-timer and possibly also someone who doesn’t mind taking it slowly. I have a full-time job so I’m unable to spend all my time working on the course.

I have no idea if this would be interesting for someone or not. I feel like perhaps there is someone out there like me who would find this an interesting opportunity. We can wait until the next version of the course is released, or alternatively we could work our way through the book (which is released and already available as a kindle version).

Anyone interested?

Just checking to see if anyone is interested in working on the book…

1 Like

Hi Alex, I’m based in Islamabad. A computer science graduate with the same routine(full time job). I’m going through the course 3 lectures a week. I’ve only completed the first 3 video chapters right now. Would love to work through this with you. How do you propose, we proceed with this?

Hi, I am currently studying software engineering in Karachi. Have just started to read the fastai book. Would love to join in on this as i am fairly new to deep learning.

Nice. Well, we could meet once every week or two and could discuss the chapters that we’re going through? There are 20 chapters, so that would talk about 6 months to go through.

How would that sound?

During the week we could post here as we are going along if there are interesting things, or problems we have? And then when we speak, someone could maybe summarise the chapter each week orally. Then we could each show what we did in terms of applying the lessons from each chapter.

I’m not sure if I’ll be able to keep to the pace of doing one chapter per week. Sometimes the chapters have a lot of new material. But definitely one chapter every two weeks would be something I could commit to.

I’m not sure how fast you’re both hoping to get through the material / course. (The video course will be out in the next week or two as well, so we can combine the two). I would rather study the material well and get a lot of practice than just rushing through the materials…

I am Byron… Kenyan

I would like to be part of the study group… No programming background
But I am always ready to learn

I have a background in Electrical and Electronics Engineering

The course notes recommend that you have at least 1 year of coding experience, so you might find it a bit hard, @Unplug_Charger. But I think there are some people who have managed it. You’re most welcome to join, though :slight_smile:, if that doesn’t put you off.

It doesn’t I am not totally green… I know pandas… Numpy… Matplotlib :sweat_smile:… Can do basic regression with Sklearn but but these I started studying around April… So I can always learn more as I progress

1 Like

Nice. I guess you should be mostly ok, then. At least enough to give it a shot? :slight_smile:

1 Like

I’ll wait to see what the others think about the timing / pace etc. @Unplug_Charger

So since I am joining 24days late you can share with me how you guys agreed to structure the six months and what you have covered so far…

We haven’t started anything yet. The other two just wrote / replied yesterday. Also, the latest version of the course hasn’t been released yet (the online version with videos), so we also probably want to wait for that. It will be released in the next week or so. So we’ll be starting from zero / scratch.

That sounds like a great plan. A chapter every week with lots of practice is comfortable enough for me. We should definitely share what we did in terms of applying the lessons from each chapter to understand the chapter more.

1 Like

@sajid22 could you please share a link to the fastai book

1 Like

Ok so to sum up for now: we’ll wait until the video course is released to start formalising this all. Luckily that will be soon.

Once it’s released, we can set up an initial meeting once we’ve all completed the first lesson etc.

Thanks all for responding, and looking forward to studying this material together.

2 Likes

Hello

So when can we start this ?

I guess now :slight_smile:

I guess we can comment here as we’re going through the first lesson. If there are questions or observations etc, we can write them here too. Then once the majority (all?) of us are finished with lesson 1, we can maybe have a call to discuss it?

In particular, though, I’d hope that we can all end up doing the practical project-based work they suggest in each lesson. I’m very curious to see what we all come up with. I’m thinking maybe this takes us somewhere between 1-2 weeks for this first lesson?

So things for our in-person discussions:

  • discuss the answers to all the questions in the ‘questionnairre’ section
  • show-and-tell of all our personal projects that we made as suggested in the course
  • any problems or issues where we learned something along the way
  • anything else…

Thats okay with me

and Alex I have some data I am preparing for collaborative filtering problem … transforming it has been a problem because my python is shallow

can I post my notebook here so that you can take a look

1 Like

I’m not that far in the course so not sure what help I can offer, but sure :slight_smile: