FastAI Bangalore Study Group Meetup
Details
First in-person meetup for FastAI Part1 (v3) Live participants in Bangalore. We'll get to know each other, cover some introductory topics, and plan for future meetups and group activities.
Agenda:
3:00-3:25 PM : Introductions
3:30-4:00 PM : Session 1 - Productivity Tips for Kaggle Competitions
4:00-4:30 PM : Session 2 - Gradient Descent and Optimization Algorithms
4:30-5:00 PM : Mixers
Session 1 - Productivity Tips for Kaggle Competitions
Speaker: Kartik Godawat, DeepKlarity
Mix of productivity tips, essential software, libraries and code structuring tips for methodically approaching ML competitions. eg:
- Get mobile notification when model training, early stopping and crashes
- Efficiently tagging models, form ensembling without confusion.
- Viewing samples from large image datasets, visualising training results etc.
Session 2 - Gradient Descent and Optimization Algorithms
Speaker: Vinod Reddy, ML Consultant
Introduction to gradient descent, the technique used for training deep learning models. The session will cover various optimization algorithms like SGD, momentum, RMSProp, Adam etc. and show sample implementation in PyTorch.
Slack group: http://dsindia.org
Location: CoWrks, RMZ Ecoworld https://goo.gl/maps/TeGrQy2t76x