Module models.rnn

Reccurent Neural Network for Shakespeare Dataset

Expand source code
#!/usr/bin/env python3

"""
Reccurent Neural Network for Shakespeare Dataset
"""

# Import PyTorch root package import torch
import torch

import torch.nn as nn


class RNN(nn.Module):

    def __init__(self, vocab_size=90, embedding_dim=8, hidden_dim=512, num_layers=2):
        super(RNN, self).__init__()

        # set class variables
        self.hidden_dim = hidden_dim
        self.vocab_size = vocab_size

        # embedding and LSTM layers
        self.embedding = nn.Embedding(vocab_size, embedding_dim)
        self.num_lstm_layers = num_layers
        self.lstm = nn.LSTM(embedding_dim, hidden_dim, batch_first=True, num_layers=num_layers)

        # linear and sigmoid layers
        self.fc = nn.Linear(hidden_dim, vocab_size)

    def forward(self, input, hidden):
        embeds = self.embedding(input)
        lstm_out, hidden = self.lstm1(embeds, hidden)
        out = self.fc(lstm_out)
        # flatten the output
        out = out.reshape(-1, self.vocab_size)
        return out, hidden

    def init_hidden(self, batch_size, device):
        hidden = (torch.zeros(self.num_lstm_layers, batch_size, self.hidden_dim).to(device = device),
                  torch.zeros(self.num_lstm_layers, batch_size, self.hidden_dim).to(device = device))
        return hidden


# TODO: Any way we can actually have an useful pretrained argument here?
def rnn(pretrained=False, num_classes=90):
    return RNN(vocab_size=num_classes)


def minirnn(pretrained=False, num_classes=90):
    return RNN(vocab_size=num_classes, hidden_dim=128)

Functions

def minirnn(pretrained=False, num_classes=90)
Expand source code
def minirnn(pretrained=False, num_classes=90):
    return RNN(vocab_size=num_classes, hidden_dim=128)
def rnn(pretrained=False, num_classes=90)
Expand source code
def rnn(pretrained=False, num_classes=90):
    return RNN(vocab_size=num_classes)

Classes

class RNN (vocab_size=90, embedding_dim=8, hidden_dim=512, num_layers=2)

Base class for all neural network modules.

Your models should also subclass this class.

Modules can also contain other Modules, allowing to nest them in a tree structure. You can assign the submodules as regular attributes::

import torch.nn as nn
import torch.nn.functional as F

class Model(nn.Module):
    def __init__(self):
        super(Model, self).__init__()
        self.conv1 = nn.Conv2d(1, 20, 5)
        self.conv2 = nn.Conv2d(20, 20, 5)

    def forward(self, x):
        x = F.relu(self.conv1(x))
        return F.relu(self.conv2(x))

Submodules assigned in this way will be registered, and will have their parameters converted too when you call :meth:to, etc.

:ivar training: Boolean represents whether this module is in training or evaluation mode. :vartype training: bool

Initializes internal Module state, shared by both nn.Module and ScriptModule.

Expand source code
class RNN(nn.Module):

    def __init__(self, vocab_size=90, embedding_dim=8, hidden_dim=512, num_layers=2):
        super(RNN, self).__init__()

        # set class variables
        self.hidden_dim = hidden_dim
        self.vocab_size = vocab_size

        # embedding and LSTM layers
        self.embedding = nn.Embedding(vocab_size, embedding_dim)
        self.num_lstm_layers = num_layers
        self.lstm = nn.LSTM(embedding_dim, hidden_dim, batch_first=True, num_layers=num_layers)

        # linear and sigmoid layers
        self.fc = nn.Linear(hidden_dim, vocab_size)

    def forward(self, input, hidden):
        embeds = self.embedding(input)
        lstm_out, hidden = self.lstm1(embeds, hidden)
        out = self.fc(lstm_out)
        # flatten the output
        out = out.reshape(-1, self.vocab_size)
        return out, hidden

    def init_hidden(self, batch_size, device):
        hidden = (torch.zeros(self.num_lstm_layers, batch_size, self.hidden_dim).to(device = device),
                  torch.zeros(self.num_lstm_layers, batch_size, self.hidden_dim).to(device = device))
        return hidden

Ancestors

  • torch.nn.modules.module.Module

Class variables

var dump_patches : bool
var training : bool

Methods

def forward(self, input, hidden) ‑> Callable[..., Any]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the :class:Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

Expand source code
def forward(self, input, hidden):
    embeds = self.embedding(input)
    lstm_out, hidden = self.lstm1(embeds, hidden)
    out = self.fc(lstm_out)
    # flatten the output
    out = out.reshape(-1, self.vocab_size)
    return out, hidden
def init_hidden(self, batch_size, device)
Expand source code
def init_hidden(self, batch_size, device):
    hidden = (torch.zeros(self.num_lstm_layers, batch_size, self.hidden_dim).to(device = device),
              torch.zeros(self.num_lstm_layers, batch_size, self.hidden_dim).to(device = device))
    return hidden