AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Machine Learning
We couldn't believe it ourselves first, right? We had an idea that it would be much better. It's actually quite complicated to calculate a number in a number that measures accuracy. If you already have a test set that is decided and is from a completely different distribution, then maybe you don't get as good results. But these are the kind of problems we had to solve and basically get around to even convince them that this is at the time.