Zero-Shot Learning
Samet Çetin
464

Hi Samet,

I’m not sure why you need to remove the last word2vec layer. Could we just use it to get the class since it can output values for each 20 classes before the softmax? I mean, that layer can possibly output zero-shot classes too. Why did you decide to compare the 300D word embedding vector output with word vectors using euclidean distance?

Thanks!

Like what you read? Give James Jaya a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.