Start on Implementing different GAN Architectures to train CNN Individually
Closed this issue · 4 comments
Ultan-Kearns commented
Start on Implementing different GAN Architectures to train CNN Individually
Ultan-Kearns commented
Got DCGAN & VAE running issue with the results and accuracy need to refactor
Ultan-Kearns commented
Having few issues optimizing the DCGANs may view more architectures, some decent results with COVID 19 radiography dataset.
Ultan-Kearns commented
DCGANs running need to work on VAE.
Ultan-Kearns commented
Seems VAE models have suffered from possible mode collapse, images seem too similar such as radiography model. Due to time / computational limitations they are to be excluded from evaluation