/Auto-encoders

Primary LanguageJupyter Notebook

Final1 file represents different variants of auto-encoders such as vanilla AE, deep AE, CAE, VAE and denoised AE and it shows images before and after compression. in Final2 file we vary the AE hyperparameters and the cost function (PSNR, SSIM, MSE, Huber). In this code, the training and test sets are concatenated into x_combined unlike the code in Final1 file. we need to note also that in Final1 file, we did some changes in AE hyperparameters everytime we create a new variant of an auto-encodernot only in Final2 file.