marvin
March 21, 2019, 5:07pm
115
Is there a plan to add CoordConv networks to fast.ai?
According to the authors:
“CoordConv models have 10-100 times fewer parameters, train in seconds rather than over an hour (150 times faster) as needed for the best-performing standard CNNs.”
Sample Pytorch implementation:
Pytorch implementation of CoordConv introduced in 'An intriguing failing of convolutional neural networks and the CoordConv solution' paper. (https://arxiv.org/pdf/1807.03247.pdf)
Paper:https://arxiv.org/pdf/1807.03247.pdf https://eng.uber.com/coordconv/
Translation invariance experiment
TL;DR: full translation invariance may not always be a good thing in time series problems. Uber’s CoordConv may be useful to help the help the model learn how much translation invariance is needed.
I’ve been intrigued by your same question @alonso in the last few months, so I decided to perform a small experiment to really test if translation invariance is always a good thing.
The main idea is very simple: can a nn learn that a sequence of 100 zeros with a 1…
2 Likes