Hi Elena. Schön zu sehen, dass noch jemand aus Deutschland ist. Aus welcher Ecke des Landes kommst du, wenn man fragen darf?
Hey Sahil. Aus welcher Ecke von Deutschland kommst du, wenn man fragen darf?
Hello folks, Shruti here. My fastai journey started in 2017. Post that, I was on and off deep learning. Not always active.
For the last couple of months I have been focused on building audio/NLP related solutions for vernacular languages. Starting with automatic speech recognition (speech to text). I took major help of @scart97.
My main goal for this year is to dive super deep in audio and NLP for non-english languages. I want to be able to extract meaningful information from the audio itself, with the smallest of data (don’t know how much of that is possible!)
Just got to know about fast_template after reading the intros above, excited to host my own blog (website?) through this.
Looking forward to diving deep into v2!
5 Likes
Thank you @jeremy! I really appreciated the first time you re-tweeted it. I re-did and cropped the animated gif and updated on github. Hopefully this is more acceptable. Here is the direct link to the gif if needed https://j.gifs.com/jZVRgW.gif.
3 Likes
Hi, everyone!
I am Leila Liu, and I am a graduate student at Tsinghua University majored in software engineering now.
It was so lucky for me to ran into fastai last year. I’ve watched the whole 2019 courses and translated the paper(UMLFit) into Chinese. To be honest, this is really useful for my successful graduation.
I’m so excited and really looking forward to attend the course this year. Thank you all!
7 Likes
Very interesting project, thanks for sharing!!
2 Likes
Great! I’d love to see the translation.
2 Likes
Hi. I’m Hoa. Checking the fastai forum is my daily habit until now is almost 2 years 
. Back in 2018, I decided to drop my PhD at the last year because I have no more motivation working on the subject and really don’t want to waste at least one more year to just get the diploma and never use that knowledge again. At that time, I contributed intensively to the fastai community and gained back a lot of knowledge about deeplearning that help me in 2 months get a job in computer vision. Since then, it’s hard to contribute that much to fastai again (especially when I just have a baby), but I’m still following up the forum everyday and write something whenever I have time or find an opportunity,
I love this community. Now it’s not just about deeplearning but where I can find motivation, encouraging peoples. Where I know I can write anything and there is someone care about it.
20 Likes
I have posted it here:
1 Like
Congrats. Having a baby was the best thing that ever happened in my life 
10 Likes
Many apologies - I remember now
Never mind 
Hi everyone,
My name is Asir. I’m from Bangladesh, currently working as an ML Research Engineer in Tokyo. I’ve been attending the fast.ai course remotely since v2.
I have a background in CS but I learned about ML/DL mostly through the fastai course. And it lead me to present a paper at the Creativity and Design Workshop at NeurIPS and the Pytorch Dev Conference. I mostly do NLP, particularly dialog systems.
I’m part of Machine Learning Tokyo, a community of ML engineers and researchers where I’ve done a few fastai study sessions. Would like to organize a remote study group with others taking the course during v4 too if anyone’s interested (and if its okay with the instructors)
I am @the_asir on twitter
3 Likes
Here you go:
Mean Field Theory of Activation Functions in Deep Neural Networks
Resurrecting the sigmoid in deep learning through dynamical…
The Emergence of Spectral Universality in Deep Networks
On the Selection of Initialization and Activation Function for…
On the Impact of the Activation Function on Deep Neural Networks Training
Dynamical Isometry and a Mean Field Theory of CNNs: How to Train…
Dynamical Isometry and a Mean Field Theory of RNNs: Gating Enables…
4 Likes
Hi everyone.
I am Julian from Argentina. My computer science career requires writing a thesis to graduate and I chose to do something related to neural networks (text detection in comics), which sounded useful and it was something I hadn’t learned in any class.
I started with standford courses which were mostly theorical and then switched over to fastai. The courses were very useful and have been using fastai since then. I’m in the last stages of the thesis so hopefully I will have time to start using fastai v2 soon, as I plan to keep doing stuff related to my hobbies and maybe try a kaggle competition some day.
I check forums frequently and its a great source of learning too, community is very nice
3 Likes
Hi everyone! I’m Edwin. I have been taking FastAI moocs from the past few years and definitely enjoy them. Definitely best DL info on the net hands down. Looking forward to the new course.
Being a developer is my day job. My Twitter @edwiiin_v
1 Like
Hello to all fastai enthusiasts
I am very grateful for again being selected, this is indeed a privilege. My home country is England, I don’t see many compatriots in this space but I am sure some are there.
I came to this space following the completion of many Coursera courses in AI. The one that pushed me here was Machine Learning Specialization from Washington University, after the 4th course Carlos Guestrin left to take up with Apple and the 5th and 6th courses were never finished.
I started with the first versions of the Moocs and look forward to the next versions each year to see what has changed.
Unfortunately because of other issues I have not been as active as I would like, hopefully that will change this time. I really enjoyed the promise of last years sessions and potential of Swift and Tensorflow, but nothing stays concrete for long, something else will best this but the thought behind it is brilliant.
At this time I have no links to offer, I don’t work for anyone but I like to play with these tools, I am perhaps the oldest enthusiast here. My work history covers many occupations, being time served as a mechanical engineer and mainly worked with my hands rather than my head until I took an Open University degree (UK) and then worked in Software specifcally in Flexible Manufacturing Systems. One of which was for delivery to Merck-Medco’s automated pharmacy in Nevada.
Anyway I note just on this page alone the diverse information presented by you world changers and know that we will succeed.
Thanks very much for the invite.
7 Likes
Oh I didn’t realize that. How were the 1st 4 courses?
The first 4 courses gave me the introduction to python and notebooks, it was also dependent on Graphlab Create if my memory serves me right. They covered Regression, Classification, and Clustering & Retrieval. The disappointing part was we did not get to the Capstone Project which was Course 6 which was to tie all the previous learning together. Although they still offer these four courses it’s probably not very popular as there is nothing to tie it all together.
1 Like