Build a Convolutional Variational AutoEncoder (CVAE) using PyTorch - Example using USPS dataset

Описание: In this video, I made a Convolutional Variational AutoEncoder (CVAE) from scratch using PyTorch. I used USPS dataset for building CVAE. This dataset consists of digit images of very low resolution (16 x 16 spatial size).

My previous video on Variational AutoEncoder (VAE):    • Build a Variational AutoEncoder (VAE) usin...  

In CVAE, I used convolutional and deconvolutional layers instead of linear layers as in VAE.

01:33 - 𝙄𝙢𝙥𝙤𝙧𝙩 𝙩𝙝𝙚 𝙡𝙞𝙗𝙧𝙖𝙧𝙞𝙚𝙨
03:35 - 𝘿𝙚𝙛𝙞𝙣𝙚 𝙩𝙝𝙚 𝙙𝙚𝙫𝙞𝙘𝙚
05:33 - 𝙇𝙤𝙖𝙙 𝙐𝙎𝙋𝙎 𝙙𝙖𝙩𝙖𝙨𝙚𝙩
08:25 - 𝘿𝙚𝙛𝙞𝙣𝙚 𝙙𝙖𝙩𝙖𝙡𝙤𝙖𝙙𝙚𝙧𝙨
09:18 - 𝘿𝙚𝙛𝙞𝙣𝙚 𝙩𝙝𝙚 𝙣𝙚𝙩𝙬𝙤𝙧𝙠 𝙖𝙧𝙘𝙝𝙞𝙩𝙚𝙘𝙩𝙪𝙧𝙚
17:26 - 𝙑𝙞𝙨𝙪𝙖𝙡𝙞𝙯𝙚 𝙩𝙝𝙚 𝙢𝙤𝙙𝙚𝙡 𝙖𝙧𝙘𝙝𝙞𝙩𝙚𝙘𝙩𝙪𝙧𝙚
18:39 - 𝙏𝙧𝙖𝙞𝙣 𝙗𝙖𝙩𝙘𝙝
20:33 - 𝙏𝙚𝙨𝙩 𝙗𝙖𝙩𝙘𝙝
22:25 - 𝙎𝙚𝙩 𝙪𝙥 𝙢𝙤𝙙𝙚𝙡 (𝘾𝙑𝘼𝙀), 𝙘𝙧𝙞𝙩𝙚𝙧𝙞𝙤𝙣 (𝙡𝙤𝙨𝙨 𝙛𝙪𝙣𝙘𝙩𝙞𝙤𝙣) 𝙖𝙣𝙙 𝙤𝙥𝙩𝙞𝙢𝙞𝙯𝙚𝙧
24:23 - 𝙏𝙧𝙖𝙞𝙣 𝙩𝙝𝙚 𝙢𝙤𝙙𝙚𝙡
31:24 - 𝙋𝙡𝙤𝙩 𝙢𝙤𝙙𝙚𝙡 𝙡𝙤𝙨𝙨
33:08 - 𝙈𝙤𝙙𝙚𝙡 𝙥𝙧𝙚𝙙𝙞𝙘𝙩𝙞𝙤𝙣
38:34 - 𝘾𝙧𝙚𝙖𝙩𝙚 𝙨𝙖𝙢𝙥𝙡𝙚𝙨 𝙛𝙧𝙤𝙢 𝙣𝙤𝙞𝙨𝙚 𝙫𝙚𝙘𝙩𝙤𝙧𝙨

GitHub address: https://github.com/randomaccess2023/M...

For mathematical details of variational autoencoders: https://stats.stackexchange.com/quest...

#data_science #deep_learning #jupyter_notebook #pytorch #convolutional_variational_autoencoder #unsupervised_learning