hello from Bhutan… since there is no deep learning group for Bhutan, Mind if I join?
Hi Guys, I’m live too…
Please add me to this group. I am Vinod and basically from Bangalore
You can join yourself on https://dsindia.herokuapp.com/
Anyone creating project groups… of 5-6 people? If anyone is interested please let me know. We can have the work discussed during the meetup. (I am new to DL).
Bit late to the party I guess. I’m Dripto. From Marathahalli, Bangalore. Nice to meet you all. It would be great to be part of the meetups and discussions.
Hi, all +1
Our next meetup is this Saturday (Nov 3): https://www.meetup.com/dsindia/events/256026299/
Do join, if you wish to interact with other folks doing FastAI Live.
Hi guys, is the live class tomorrow at 8am?
The live session class starts at 7 am
Did you receive any link for that?
No, I have not received the link yet
As per Google calendar shared, it is at 8 tomorrow. No link yet
Due to daylight savings in time, its delayed for us by an hour
First time here. I am from Kolkata, currently in Bangalore. Just started this course yesterday and had a question.
I have seen many a times on Reddit and other websites that people make use of CNNs and make emotion detectors which take live photos via the webcam and predict the emotions?
Sorry for being naive as this is my first DL course. But my vague idea of doing this is same like I had done some ML problems before.
- Download a face emotion dataset from Google.
- Train a CNN like this lesson made me learn.
- Pass the webcame images into this and predict results out.
This is just an idea. Currently I am stuck at the third step. There was a thing called validation set which, as was mentioned, the algorithm does not see. Is that the set where I pass these webcam images to be predicited results upon?
I really want to work upon this. If anyone can help, please reply to this. Heard great things about the fast.ai forums.
Welcome to fastai. Your queries related to the data and the label in FASTAI is available here https://docs.fast.ai/vision.data.html.
Basically you can provide data as TRAIN VALID and TEST. If your valid set is without the labels then treat that folder as TEST actually here, and assign 20% of your training dataset as your validation set.
I will definitely check the link out and contact you for further questions
Hello, I’ve just launched a discussion thread on fastai study groups to gather feedbacks from the organizers and participants in order to list the best practices and also to avoid some gaps.
You will find more information in this post. Thank you if you can take a few minutes to participate in the discussion