AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Limit in Machine Learning
I remember something about you being able to identify 30, 40 features in the imagery as it was then. And I realize now that you've increased that to 70 80. Is this just because of the new camera? Is it a better model? What's changed? Like what made that change happen?What we did were up to our fifth generation of our system now,. The fifth generation has I think 78 layers in it. It's actually the same model that runs in Australia, New Zealand, Canada and the US. So it's this really cool, elegant solution of one model, but many outputs.