Welcome to Part 1 (v2)

(sergii makarevych) #21

A story in pictures :slight_smile:

(Sanyam Bhutani) #22

May I suggest adding link to the ‘how has your journey been’ thread to the post as well?

(Jeremy Howard (Admin)) #23

Good idea.

(Sanyam Bhutani) #24

@Jeremy, I’ve another suggestion: nurture.ai is hosting Saturday meet ups and they’d be discussing Fast AI course during the meet ups as well- you can link it here to help people seeking communities or cohorts to learn together with (They are running sessions across all the major cities)


(George G.) #25

Thanks a lot Jeremy! I was waiting for the second release a lot. George

(George G.) #26

In addition to David Silver’s and Berkeley’s AI there is a new course by MSFT and EDX https://courses.edx.org/courses/course-v1:Microsoft+DAT257x+3T2017/course/. They have some problems with course’s internal website… for some reason the site is not working. You need to go to Discussions tab where they pinned one post about how to reach each module :slight_smile: and start from there. You have a lot of labs and course is really good.

Then there is another new RL MOOC on coursera, https://www.coursera.org/learn/practical-rl, which starts now and which will have some interesting and tough PAs (at least in other courses of this AML specialization they did have good programming assignments).

Finally Berkeley has the class that combines Deep Learning (mainly for learning representation and RL and they publish videos every semester with programming assignments and everything. Just google it.

(Jeremy Howard (Admin)) #27

@init_27 do you know if they’ll be studying the new course, or the old one?

(Sanyam Bhutani) #28

@jeremy We were waiting for the public release (I’m a volunteer and will be helping in organising the meetups in my city).

So v2, if it’s okay with you? Or should we wait for the final release? The community has grown to 2k+ members in 2 weeks-so I wasn’t sure if I should share about the release with everyone yet because it’s a soft release.

(Jeremy Howard (Admin)) #29

Definitely should be v2!

(Sanyam Bhutani) #30


We’ll be starting with the v2 meets on every ‘AI Saturday’ meetup from 6th of January onwards. (A lot of the ‘Ambassadors’ or organisers are international fellows)

As of now, there’s a confirmation of meetups (on this Saturday) in Singapore, Hyderabad, Bangalore, New Delhi, Kuala Lumpur, Karachi, Los Angeles, Guwahati, Jakarta.

We hope to bring the meetups to full throttle across all 61 cities very soon.

(Seema Goel) #31

I am from Bangalore too, the link is showing inactive to me.

(Sudhanshu Garg) #32

Hi… yes it expired. Try this.

(Seema Goel) #33

Thanks It worked

(sergii makarevych) #34

I am a bit confused - was the 1-st lesson changed?

(Jeremy Howard (Admin)) #35

Yes I re-recorded lesson 1.


So this course is no longer of any use then http://course.fast.ai/

‘Practical Deep Learning for Coders Part-1’

(Sanyam Bhutani) #37

It’s not completely useless. But you’d find the v2 more (much much more) useful than the v1. It’s highly recommended you follow the v2 of the course :slight_smile:


That’s Okay.Was just thinking of starting to do the v1 course on my own.

Since it’s such a well know course,thought i should do it as it will help me learn and understand deep learning more than what i know currently.

(Jeremy K) #39

So what exactly is the difference? Is it like more information?

(Ahmad Arib) #40

@jeremy they will study the new one, part 1 v2.
Below I copy the notification from #general workspace on Slack
@everyone fast.ai version 2 (Welcome to Part 1 (v2)) was released as a low-profile soft launch yesterday morning, and we have decided to move to fast.ai v2 for the Practical Deep Learning sessions. Easy-to-follow environment setup instructions for Crestle and Paperspace are available in Lecture 1: https://www.youtube.com/watch?v=IPBSB1HLNLo&feature=youtu.be

The AI Saturdays team is working on creating easy-to-follow detailed instructions on how to set-up for fast.ai v2 on Google Cloud Platform, and will get back shortly once we have that