Variational Autoencoders for MNIST Dataset
I reduced the running time from 681 seconds to 360 seconds for 50 training epochs. The code was optimized by processing the raw data and storing it as two big tensors in the .pt format.
Created on January 6|Last edited on January 6
Comment
Graphical Results
reconstructed
generated
original
Run: jolly-glade-4
1
Add a comment