Member-only story
Basically what an auto-encoder does is it takes some kind of input data it could an image or a vector anything at all with a very high dimensionality. it is going pass these data in a neural network and it gonna try and compress the data into a smaller representation it does this with two principal components is what we call the encoder .
- Autoencoder takes input data it could be Image or vector with high Dimensionality.
- It gonna try and compress the data into a smaller representation it does this with two principal components is what we call encoder .
- From the Latent space with less dimension , the network will try to reconstruct the input by using again convolutional layer.
- Loss function is computed by comparing input to output with the pixel difference.
This is simply a bunch of layers they could be fully connected layer or convolution layer that are going to take the input and they are going to compress it down to a smaller representation. which has less dimension that of input , from he bottle neck is going to try and reconstruct the input by using again fully connected or convolution layers.
