VariationalEncoder

source

VariationalEncoder(
   model_params, dataset_params: Dict[str, int]
)

---

Defining the architecture of a MLP encoder

Args

• model_params (dict) : dictionary containing the model parameters
• dataset_params (dict) : dictionary containing the dataset parameters

Attributes

• seq_len (int) : length of the input sequence
• alphabet_size (int) : size of the alphabet
• hidden_layer_sizes (list) : list containing the sizes of the hidden layers
• z_dim (int) : dimension of the latent space
• dropout_prob (float) : dropout probability
• non_linearity (function) : non-linearity function

Methods:

.forward

source

.forward(
   x: torch.Tensor
)