Introduce yourself here

Edward · February 26, 2020, 8:25pm

Thanks for the brilliant work on deOldify. I’ve had fun deOldifying images for a couple of local Facebook groups that publish old photos of the local town and village. Some of them turned out quite stunning. I was wondering if one could specialise the model using localised photos with more of the “correct” colours for as in this case Swedish buildings, cars and such?

muellerzr · February 27, 2020, 5:51pm

Just thought I would briefly mention this here, if anyone is planning on doing the course via Google Colaboratory, I work out of it exclusively and have been using it for fastai2 since the dev version was first released. If you have issues getting it working feel free to @ me (Also once the course comes out I’ll write a How-To on setting up Colab)

liberus · February 27, 2020, 11:01pm

Can you say how easy/frustrating it is to use colab for fastai2? This might help study group I’m planning to organize.

muellerzr · February 27, 2020, 11:02pm

Significantly easy

Bhuvana_ka · March 3, 2020, 2:50pm

Hey All,

I am Bhuvana from Chennai, India. An edtech entrepreneur turned ML researcher. Currently am working as Research engineer for a Data analytics company ( clinical trials domain) .

I started learning Deep Learning with fast.ai part 2 online videos while i was running my edtech company founded in 2013… The courses gave me the foundation and confidence. I wound up my company after 6 long years and joined as an intern with a Data analytics company mid of last year where i currently work as an ML researcher.

Thanks to Jeremy, Rachel , Sylvain, team and the folks in the discussion forums, am working as an ML researcher. Thanks to Fast ai, I have gone from a novice to researcher in an year.

I want to take a deep dive and develop solutions to problems in the clinical trials domain.

Thanks fast ai team for this opportunity and looking forward to the course.

my twitter - @KA_Bhuvana

jianshen92 · March 4, 2020, 1:17am

Hi All,

I am Jian Shen from Kuala Lumpur, Malaysia. Currently, I am working as a data scientist in AirAsia (Malaysian low-cost airline). The airline business is rather slow recently due to the Covid-19 outbreak

I started my career as a Software Engineer. Two years into the field, I realised that I wanted to augment my skillsets in the domain of Data Science/Machine Learning to differentiate myself among other software engineers, and Mathematics is always my forte growing up. I made a switch and joined a startup as ML Engineer, that is where I first got my hands on experience in Machine Learning.

Then Fast.ai found me, that’s where I managed to get shortlisted in an AI Challenge hosted by Grab (South East Asia equivalent of Uber/Lyst) by using fast.ai to solve a Computer Vision problem. Yay!

I aspire to become a researcher in this field and I really enjoy the experimental mindset of the fast.ai community! Looking forward for v2 release!

ecdrid · March 5, 2020, 5:02am

Hi Everyone, I am from India, currently working as ML Engineer @TCS in Bangalore;

I first experimented with machine learning in late 2016. Since then I have done the CNN course by Stanford(CS-231N, CS-224D), Machine Learning by Andrew NG(CS-229), CMU’s DL, Machine Learning course by Jeremy and his Team, Fast.ai (ML + DL ) and other online courses from Coursera, ODS etc. Going through Kaggle kernels, Medium, SO posts and competing on Kaggle comps is part of my daily job now…

In short , completely self taught from wonderful resources available online…

Had been the part of Fast.ai International fellowship ever since the team switched to PyTorch; Was lucky enough to get in again!

I won’t lie, couldn’t focus much in the last run of the FastAI course due to the job in hand but being invited again this time I am gonna do justice to the course! Thanks a lot for that Jeremy!

By the way there are many names I can recognise from previous runs!
Good to see you guys again! Looking forward to learn a ton of things from the legends again!

And always remember, Doing A Course isn't enough, We need to Apply the Learning!

Best Regards,
Aditya.(@ecdrid)

nok · March 7, 2020, 12:51pm

Hi I am Nok from Hong Kong. Learning new things from fastai every year and looking forward to it, I hope to contribute more this year. Would love to join study group in similar timezone.

utkb · March 10, 2020, 6:59pm

Hi all. My name is Yijin, currently based in Birmingham, UK. I started learning about and using fastai during the first PyTorch-based version of the “Part I” course in 2018, before the pre-release (aka v0.7 : ) of fastai library morphed into v1. I have also followed different iterations of “Part II” material.

Since then, I have used the fastai v1 library to explore different fields within my work at engineering consultancy firm Arup. For example:

Tree species identification via tree bark image classification, for environmental conservation. Proof-of-concept drew on our own locale-specific data and also open-source data like the nicely organised BarkNet.

Automated text generation of preliminary project info (e.g. context, aims, value, methods, etc.) from database of previous projects and basic user input prompting, using the very adaptable and user-friendly workflow of fastai’s ULMFiT.

Ordinal image classification of seismic structural damage in PEER Hub ImageNet Challenge, specifically “Scene Level” followed by “Damage State” (task descriptions here).

From fastai, when I was exploring SwiftAI and S4TF, I came across this excellent post by James Thompson, and for entertainment I applied a classic (or I guess clichéd? =P ) style transfer of Van Gogh’s Starry Night onto a photo of company founder Sir Ove Arup (yeah, sometimes we are a bit like a cult…)

StarryOve800px800×800 532 KB

ove_arup_head_800px800×800 174 KB

In the past few months I have not been following fastai and the forum as closely as I would have liked, having been busy helping with the next release of the CAE software from Arup’s software house (shameless plug: Oasys LS-DYNA software suite), among other R&D work and some travelling. And so it was a great surprise when realised that I have been given the privilege of an invite to the 2020 course livestream here! : )

I am trying to last-minute sort out a virtual (sigh, covid19…) study group for a few participants within my company/team, and no matter what I will definitely schedule my time to follow the livestreams and the much-anticipated fastbook. Hopefully this ‘renewed experience’ will follow the same exciting route that I had — previously fastai v0.7 to v1 applications, and now v2 pre-release to future applications.

Looking forward to all the great learning and discussions ahead! Thanks to all.

Yijin

magiclantern · March 10, 2020, 10:32pm

Hi all! Seth here.

I’ve previously completed Practical Deep Learning for coders, part 1 and most of part 2 in 2019. Really enjoyed the courses and got some deep learning on deep learning. Have had the chance to do some tabular, text, and time series deep learning since then as part of my work. My deep learning interests are primarily with text and time series data. While I have not done any work myself with imaging, I would like to be familiar with the techniques and know how to apply them.

My educational background covers Computer Science, Business Administration, and Biomedical Informatics. I’ve worked in Higher Ed., E-Commerce, and for the last 12 years or so been in the Health Care space.

Currently located in Colorado, USA.

Follow me on twitter @magic__lantern

nok · March 11, 2020, 2:06am

that’s so cool to have study group within your team.

Brad_S · March 11, 2020, 5:19am

Not the right thread… But do ipywidgets work there now?

muellerzr · March 11, 2020, 5:20am

Some do some don’t. But ImageCleaner is unsupported, even Jeremy’s new one (Look at my ClassConfusion widget from fastaiv1 to see an example, both my own Colab version and the one in the fastai library for ideas on migrating and integrating the two)

asparagui · March 11, 2020, 10:17pm

hello fastai world! i have been working a lot with swift for tensorflow this past year. more about me/some of my work: http://brettkoonce.com/talks/
happy to chat about whatever, feel free to reach out! looking forward to the new course and fastai v2 in general!

init_27 · March 12, 2020, 3:12pm

Hello!
Humble request to the great mobile experts, @asparagui & @machinethink. If you have the time and if you might be interested, please do share some resources or share your great insights with us in a small study group

Thanks so much in advance!

abhinow · March 12, 2020, 4:18pm

Hi everyone, Super excited to be a part of this and I look forward to the lectures. Would also love to join a study group that may be available. Thanks.

nissan · March 12, 2020, 5:49pm

Hi everyone,
My name is Nissan Dookeran and I’m originally from Trinidad and Tobago, but now living in Brisbane, Australia. I’m very excited to join the 2020 course for Deep Learning 1 as the “formal education” part of my Deep Learning journey. I previously participated in the MOOC for v3 of Deep Learning 1, and even built a hummingbird classifier that @jeremy mentioned briefly in one of his lectures from DL1 last year. I’m professionally a full-stack developer but am hoping with more formal qualifications I can finally transition my passionate “hobby” for this field into a professional career transition by deepening my understanding and readings of research into the latest developments of the field, and hopefully contributing back to the open source community in some way with what I learn. Thank you again to the organisers for creating this course and workshops and making them so accessible, I never would have found this new passion and definitely not have the level of understanding I currently do without it. If there are any Australia-based persons looking to start a study group please let me know, I’m very keen. My twitter handles are @nissandookeran (personal/pro tweets) and @redditech (strictly professional tweets)