This was a great explanation, however I don’t understand why at we don’t want to do the reparameterization trick during testing and only return the mu? I would assume we would want to always sample from the latent distribution for passing it to the decoder? Making the encoder give deterministic output (just the mu) during testing will defeat the purpose of variational auto encoders right?(00:52:00 - 00:58:05)
Week 8 – Practicum: Variational autoencoders
Alfredo Canziani
※本サイトに掲載されているチャンネル情報や動画情報はYouTube公式のAPIを使って取得・表示しています。
Timetable
動画タイムテーブル