| Autoencoder | Notebook | ckpts | MSE |
|---|---|---|---|
| Vanilla autoencoder | .ipynb | 🔗 | 0.014516565627219851 |
| Deep Convolutional autoencoder | .ipynb | 🔗 | 0.00950322496098459 |
| Denoising Autoencoder | .ipynb | 🔗 | 0.009893021930423705 |
For all autoencoders, I have used tensorflow, keras and MNIST Dataset
git clone https://github.com/shoryasethia/Autoencoder
cd Autoencoder
Autoencoders are a type of neural network used for unsupervised learning. Autoencoders consist of an encoder which processes the input into some latent features and a decoder which uses those latent features and tries to reconstruct the original image, they work together to learn an efficient representation of the input data.
The architecture of an autoencoder typically consists of an input layer, a hidden layers (encoder), and an output layer(decoder). The encoder compresses the input data into a latent space representation, while the decoder reconstructs the original input from the latent space representation.
- If this repo helped in any way, star this repo.
- Author : @shoryasethia