models.py

import torch
from torch import nn
import torchvision

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

class Encoder(nn.Module):
    '''Encoder Model'''

    def __init__(self, encoded_image_size = 14):
        super(Encoder, self).__init__()
        self.enc_image_size = encoded_image_size

        # pretrained ImageNet ResNet-101
        resnet = torchvision.models.resnet101(pretrained=True)

        # Remove linear and pool layers (since we're not doing classification)
        modules = list(resnet.children())[:-2]
        self.resnet = nn.Sequential(*modules)

        # Resize the image to a fixed size to allow input images of variable size
        self.adaptive_pool = nn.AdaptiveAvgPool2d((encoded_image_size,encoded_image_size))

        self.fine_tune()


    def forward(self, images):
        '''
        Forward propagation
        :param images: images, a tensor of dimensions (batch_size, 3, image_size, image_size)
        :return: encoded images
        '''

        out = self.resnet(images) # (batch_size, 2048, image_size/32, image_size/32)
        out = self.adaptive_pool(out) # (batch_size, 2048, encoded_image_size, encoded_image_size)
        out = out.permute(0,2,3,1) # (batch_size, encoded_image_size, encoded_image_size, 2048)
        return out

    def fine_tune(self, fine_tune=True):
        '''
        Allow or prevent the computation of gradients for convolutional blocks 2 through 4 of the encoder.
        :param fine_tune: Allow?
        '''

        for p in self.resnet.parameters():
            p.requires_grad = False

        for c in list(self.resnet.children())[5:]:
            for p in c.parameters():
                p.requires_grad = fine_tune

    
class Attention(nn.Module):
    '''Attention Network'''

    def __init__(self, encoder_dim, decoder_dim, attention_dim):
        """
        :param encoder_dim: feature size of encoded images
        :param decoder_dim: size of decoder's RNN
        :param attention_dim: size of the attention network
        """
        super(Attention, self).__init__()
        # linear layer to transform encoded image
        self.encoder_att = nn.Linear(encoder_dim, attention_dim)  
        # linear layer to transform decoder's output
        self.decoder_att = nn.Linear(decoder_dim, attention_dim)
        # linear layer to calculate values to be softmax-ed
        self.full_att = nn.Linear(attention_dim,1) 
        self.relu = nn.ReLU()
        # softmax layer to calculate weights
        self.softmax = nn.Softmax(dim=1)

    def forward(self, encoder_out, decoder_hidden):
        """
        Forward propagation.
        :param encoder_out: encoded images, a tensor of dimension (batch_size, num_pixels, encoder_dim)
        :param decoder_hidden: previous decoder output, a tensor of dimension (batch_size, decoder_dim)
        :return: attention weighted encoding, weights
        """
        att1 = self.encoder_att(encoder_out)  # (batch_size, num_pixels, attention_dim)
        att2 = self.decoder_att(decoder_hidden) # (batch_size, attention_dim)
        att = self.full_att(self.relu(att1 + att2.unsqueeze(1))).squeeze(2)  # (batch_size, num_pixels)
        alpha = self.softmax(att) # (batch_size, num_pixels)
        attention_weighted_encoding = (encoder_out * alpha.unsqueeze(2)).sum(dim=1)  # (batch_size, encoder_dim)
        
        return attention_weighted_encoding, alpha


class DecoderWithAttention(nn.Module):
    '''Decoder'''
    def __init__(self, attention_dim, embed_dim, decoder_dim, vocab_size, encoder_dim=2048, dropout=0.5):
        '''
        :param attention_dim: size of attention network
        :param embed_dim: embedding size
        :param decoder_dim: size of decoder's RNN
        :param vocab_size: size of the vocabulary
        :param encoder_dim: feature size of encoded images
        :param dropout: dropout
        '''
        super(DecoderWithAttention, self).__init__()

        self.encoder_dim = encoder_dim  #2048
        self.attention_dim = attention_dim   #512
        self.embed_dim = embed_dim  #512
        self.decoder_dim = decoder_dim  # 512
        self.vocab_size = vocab_size
        self.dropout = dropout

        # Create an Attention Network Instance
        self.attention = Attention(encoder_dim, decoder_dim, attention_dim)
        
        # Embedding Layer
        self.embedding = nn.Embedding(vocab_size, embed_dim) 
        # Dropout Layer 
        self.dropout = nn.Dropout(p = self.dropout)
        # Decoding LSTM
        self.decode_step = nn.LSTMCell(embed_dim + encoder_dim, decoder_dim,bias=True)
        # linear layer to find initial hidden state of LSTMCell
        self.init_h = nn.Linear(encoder_dim, decoder_dim)
        # linear layer to find initial cell state of LSTMCell
        self.init_c = nn.Linear(encoder_dim, decoder_dim)
        # linear layer to create a sigmoid activated gate
        self.f_beta = nn.Linear(decoder_dim, encoder_dim)
        self.sigmoid = nn.Sigmoid()
        # linear layer to find scores over vocab
        self.fc = nn.Linear(decoder_dim, vocab_size)
        # initialize some layers with the uniform distribution
        self.init_weights()


    def init_weights(self):
        '''
        Initializes some parameters with values from the uniform distribution, for easier convergence
        '''
        self.embedding.weight.data.uniform_(-0.1,0.1)
        self.fc.bias.data.fill_(0)
        self.fc.weight.data.uniform_(-0.1, 0.1)


    def load_pretrained_embeddings(self, embeddings):
        """
        Loads embedding layer with pre-trained embeddings.
        :param embeddings: pre-trained embeddings
        """
        self.embedding.weight = nn.Parameter(embeddings)


    def fine_tune_embeddigs(self, fine_tune=True):
        """
        Allow fine-tuning of embedding layer? (Only makes sense to not-allow if using pre-trained embeddings).
        :param fine_tune: Allow?
        """
        for p in self.embedding.parameters():
            p.requires_grad = fine_tune

    
    def init_hidden_state(self, encoder_out):
        """
        Creates the initial hidden and cell states for the decoder's LSTM based on the encoded images.
        :param encoder_out: encoded images, a tensor of dimension (batch_size, num_pixels, encoder_dim)
        :return: hidden state, cell state
        """
        mean_encoder_out = encoder_out.mean(dim=1)
        # mean_encoder_out = [batch_size, encoder_dim]
        h = self.init_h(mean_encoder_out)  # (batch_size, decoder_dim)
        c = self.init_c(mean_encoder_out)  
        return h, c


    def forward(self, encoder_out, encoded_captions ,caption_lengths):
        '''
        Forward Propagation

        :param encoder_out: encoded images, a tensor of dimention (batch_size,enc_image_size, enc_image_size, encoder_dim)
        :param encoded_captions: encoded captions,  a tensor of dimension (batch_size, max_caption_length)
        :param caption_lengths: caption lengths, a tensor of dimension (batch_size, 1)
        :return: scores for vocabulary, sorted encoded captions, decode lengths, weights, sort indices
        '''

        batch_size = encoder_out.size(0)
        encoder_dim = encoder_out.size(-1) # 2048
        vocab_size = self.vocab_size

        # encoder_out.shape = (batch_size, 14,14,2048)
        # Flatten the image:  
        encoder_out= encoder_out.view(batch_size, -1, encoder_dim)
        # encoder_out.shape = (batch_size, 14*14,2048)
        num_pixels = encoder_out.size(1)

        # Sort input data by decreasing lengths:
        caption_lengths, sort_ind = caption_lengths.squeeze(1).sort(dim=0, descending=True)
        # caption_lengths = sorted([batch_size]), sort_ind = [batch_size]
        encoder_out = encoder_out[sort_ind]
        encoded_captions = encoded_captions[sort_ind]

        # Embedding
        embeddings = self.embedding(encoded_captions) 
        # embedding.shape = [batch_size, max_caption_length, embed_dim]

        # Initialize the LSTM state
        h, c = self.init_hidden_state(encoder_out) # (batch_size, decoder_dim)

        # We won't decode at the <end> position, since we've finished generating as soon as we generate <end>
        # so, decoding lengths are actual lengths-1
        decode_lengths = (caption_lengths -1 ).tolist()
        # decode_lengths = [batch_size]

        # create tensors to hold word prediction scores and alphas
        predictions = torch.zeros(batch_size, max(decode_lengths), vocab_size).to(device)
        alphas = torch.zeros(batch_size, max(decode_lengths), num_pixels).to(device)

        # At each time-step, decode by 
        # attention-weighing the encoder's output based on the decoder's based on the decoders previous hidden state output
        # then generate a new word in the decoder with the previous word and the attention weighted encoding
        for t in range(max(decode_lengths)):
            batch_size_t = sum([ l > t for l in decode_lengths])
            attention_weighted_encoding, alpha = self.attention(encoder_out[:batch_size_t], h[:batch_size_t])

            gate = self.sigmoid(self.f_beta(h[:batch_size_t])) # gating scalar, (batch_size_t, encoder_dim)
            attention_weighted_encoding = gate * attention_weighted_encoding
            h, c = self.decode_step(
                torch.cat([embeddings[:batch_size_t, t, :], attention_weighted_encoding], dim=1),
                (h[:batch_size_t], c[:batch_size_t]))  # (batch_size_t, decoder_dim)

            preds = self.fc(self.dropout(h)) # (batch_size_t, vocab_size)
            predictions[:batch_size_t, t, :] = preds
            alphas[:batch_size_t, t, :] = alpha
        
        return predictions, encoded_captions, decode_lengths, alphas, sort_ind