Python torch.nn() Examples

def __init__(self):
        super(Model2, self).__init__()

        # fine tuning the ResNet helped significantly with the accuracy
        base_model = MyResNet(BasicBlock, [2, 2, 2, 2])
        base_model.load_state_dict(model_zoo.load_url(model_urls['resnet18']))
        # code needed to deactivate fine tuning of resnet
        #for param in base_model.parameters():
        #    param.requires_grad = False
        self.base_model = base_model
        self.drop0 = nn.Dropout2d(0.05)

        self.conv1 = nn.Conv2d(512, 256, 3, padding=1, bias=False)
        self.bn1 = nn.BatchNorm2d(256)
        self.drop1 = nn.Dropout2d(0.05)

        self.conv2 = nn.Conv2d(256, 128, 3, padding=1, bias=False)
        self.bn2 = nn.BatchNorm2d(128)
        self.drop2 = nn.Dropout2d(0.05)

        self.conv3 = nn.Conv2d(128, 1+9, 3, padding=1, bias=False) 

def __init__(self, block, layers, num_classes=1000):
        self.inplanes = 64
        super(MyResNet, self).__init__()
        self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3,
                               bias=False)
        self.bn1 = nn.BatchNorm2d(64)
        self.relu = nn.ReLU(inplace=True)
        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
        # note the increasing dilation
        self.layer1 = self._make_layer(block, 64, layers[0])
        self.layer2 = self._make_layer(block, 128, layers[1], stride=2, dilation=1)
        self.layer3 = self._make_layer(block, 256, layers[2], stride=1, dilation=2)
        self.layer4 = self._make_layer(block, 512, layers[3], stride=1, dilation=4)

        # these layers will not be used
        self.avgpool = nn.AvgPool2d(7)
        self.fc = nn.Linear(512 * block.expansion, num_classes)

        for m in self.modules():
            if isinstance(m, nn.Conv2d):
                n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
                m.weight.data.normal_(0, math.sqrt(2. / n))
            elif isinstance(m, nn.BatchNorm2d):
                m.weight.data.fill_(1)
                m.bias.data.zero_() 

def __init__(self, block, layers, num_classes=1000):
    self.inplanes = 64
    super(ResNet, self).__init__()
    self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3,
                 bias=False)
    self.bn1 = nn.BatchNorm2d(64)
    self.relu = nn.ReLU(inplace=True)
    # maxpool different from pytorch-resnet, to match tf-faster-rcnn
    self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
    self.layer1 = self._make_layer(block, 64, layers[0])
    self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
    self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
    # use stride 1 for the last conv4 layer (same as tf-faster-rcnn)
    self.layer4 = self._make_layer(block, 512, layers[3], stride=1)

    for m in self.modules():
      if isinstance(m, nn.Conv2d):
        n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
        m.weight.data.normal_(0, math.sqrt(2. / n))
      elif isinstance(m, nn.BatchNorm2d):
        m.weight.data.fill_(1)
        m.bias.data.zero_() 

def __init__(self, input_size, hidden_size, correlation_func=1, do_similarity=False):
        super(AttentionScore, self).__init__()
        self.correlation_func = correlation_func
        self.hidden_size = hidden_size

        if correlation_func == 2 or correlation_func == 3:
            self.linear = nn.Linear(input_size, hidden_size, bias=False)
            if do_similarity:
                self.diagonal = Parameter(torch.ones(1, 1, 1) / (hidden_size ** 0.5), requires_grad=False)
            else:
                self.diagonal = Parameter(torch.ones(1, 1, hidden_size), requires_grad=True)

        if correlation_func == 4:
            self.linear = nn.Linear(input_size, input_size, bias=False)

        if correlation_func == 5:
            self.linear = nn.Linear(input_size, hidden_size, bias=False) 

def _init_modules(self): 
    self._init_head_tail()

    # rpn
    self.rpn_net = nn.Conv2d(self._net_conv_channels, cfg.RPN_CHANNELS, [3, 3], padding=1)

    self.rpn_cls_score_net = nn.Conv2d(cfg.RPN_CHANNELS, self._num_anchors * 2, [1, 1])
    
    self.rpn_bbox_pred_net = nn.Conv2d(cfg.RPN_CHANNELS, self._num_anchors * 4, [1, 1])

    self.cls_score_net_fast = nn.Linear(self._fc7_channels, self._num_classes+1)
    self.bbox_pred_net_fast = nn.Linear(self._fc7_channels, (self._num_classes+1) * 4)


    self.cls_score_net = nn.Linear(self._fc7_channels, self._num_classes)  # between class
    self.bbox_pred_net = nn.Linear(self._fc7_channels, self._num_classes)  # between boxes

    self.init_weights() 

def make_model(d_vocab, N, d_model, d_ff=1024, h=4, dropout=0.1):
    """Helper: Construct a model from hyperparameters."""
    c = copy.deepcopy
    attn = MultiHeadedAttention(h, d_model)
    ff = PositionwiseFeedForward(d_model, d_ff, dropout)
    position = PositionalEncoding(d_model, dropout)
    model = EncoderDecoder(
        Encoder(EncoderLayer(d_model, c(attn), c(ff), dropout), N),
        Decoder(DecoderLayer(d_model, c(attn), c(attn), c(ff), dropout), N),
        nn.GRU(d_model, d_model, 1),
        nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        Generator(d_model, d_vocab),
        d_model
    )
    # This was important from their code.
    # Initialize parameters with Glorot / fan_avg.
    for p in model.parameters():
        if p.dim() > 1:
            nn.init.xavier_uniform_(p)
    return model 

def _make_layer(self, block, planes, blocks, stride=1):
    downsample = None
    if stride != 1 or self.inplanes != planes * block.expansion:
      downsample = nn.Sequential(
        nn.Conv2d(self.inplanes, planes * block.expansion,
              kernel_size=1, stride=stride, bias=False),
        nn.BatchNorm2d(planes * block.expansion),
      )

    layers = []
    layers.append(block(self.inplanes, planes, stride, downsample))
    self.inplanes = planes * block.expansion
    for i in range(1, blocks):
      layers.append(block(self.inplanes, planes))

    return nn.Sequential(*layers) 

def make_model(d_vocab, N, d_model, latent_size, d_ff=1024, h=4, dropout=0.1):
    """Helper: Construct a model from hyperparameters."""
    c = copy.deepcopy
    attn = MultiHeadedAttention(h, d_model)
    ff = PositionwiseFeedForward(d_model, d_ff, dropout)
    position = PositionalEncoding(d_model, dropout)
    share_embedding = Embeddings(d_model, d_vocab)
    model = EncoderDecoder(
        Encoder(EncoderLayer(d_model, c(attn), c(ff), dropout), N),
        Decoder(DecoderLayer(d_model, c(attn), c(attn), c(ff), dropout), N),
        # nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        # nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        nn.Sequential(share_embedding, c(position)),
        nn.Sequential(share_embedding, c(position)),
        Generator(d_model, d_vocab),
        c(position),
        d_model,
        latent_size,
    )
    # This was important from their code.
    # Initialize parameters with Glorot / fan_avg.
    for p in model.parameters():
        if p.dim() > 1:
            nn.init.xavier_uniform_(p)
    return model 

def __init__(self):
        super(Model, self).__init__()
        self.conv1 = nn.Conv2d(3, 16, 3, padding=1)
        self.bn1 = nn.BatchNorm2d(16)

        self.conv2 = nn.Conv2d(16, 32, 3, padding=1)
        self.bn2 = nn.BatchNorm2d(32)

        self.conv3 = nn.Conv2d(32, 64, 3, padding=1)
        self.bn3 = nn.BatchNorm2d(64)

        self.conv4 = nn.Conv2d(64, 128, 3, padding=1)
        self.bn4 = nn.BatchNorm2d(128)

        self.conv5 = nn.Conv2d(128, 128, 3, dilation=2, padding=2)
        self.bn5 = nn.BatchNorm2d(128)

        self.conv6 = nn.Conv2d(128, 128, 3, dilation=4, padding=4)
        self.bn6 = nn.BatchNorm2d(128)

        self.conv7 = nn.Conv2d(128, 1+9, 3, padding=1) 

def make_model(d_vocab, N, d_model, d_ff=1024, h=4, dropout=0.1):
    """Helper: Construct a model from hyperparameters."""
    c = copy.deepcopy
    attn = MultiHeadedAttention(h, d_model)
    ff = PositionwiseFeedForward(d_model, d_ff, dropout)
    position = PositionalEncoding(d_model, dropout)
    model = EncoderDecoder(
        Encoder(EncoderLayer(d_model, c(attn), c(ff), dropout), N),
        Decoder(DecoderLayer(d_model, c(attn), c(attn), c(ff), dropout), N),
        nn.GRU(d_model, d_model, 1),
        nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        Generator(d_model, d_vocab),
        d_model
    )
    # This was important from their code.
    # Initialize parameters with Glorot / fan_avg.
    for p in model.parameters():
        if p.dim() > 1:
            nn.init.xavier_uniform_(p)
    return model 

def __init__(self, d_model, vocab):
        super(Generator, self).__init__()
        self.proj = nn.Linear(d_model, vocab) 

def clones(module, N):
    """Produce N identical layers."""
    return nn.ModuleList([copy.deepcopy(module) for _ in range(N)]) 

def __init__(self, latent_size, output_size):
        super().__init__()
        self.fc1 = nn.Linear(latent_size, 100)
        self.relu1 = nn.LeakyReLU(0.2, )
        self.fc2 = nn.Linear(100, 50)
        self.relu2 = nn.LeakyReLU(0.2)
        self.fc3 = nn.Linear(50, output_size)
        self.sigmoid = nn.Sigmoid() 

def __init__(self, size, dropout):
        super(SublayerConnection, self).__init__()
        self.norm = LayerNorm(size)
        self.dropout = nn.Dropout(dropout) 

def __init__(self, features, eps=1e-6):
        super(LayerNorm, self).__init__()
        self.a_2 = nn.Parameter(torch.ones(features))
        self.b_2 = nn.Parameter(torch.zeros(features))
        self.eps = eps 

def __init__(self, d_model, dropout, max_len=5000):
        super(PositionalEncoding, self).__init__()
        self.dropout = nn.Dropout(p=dropout)

        # Compute the positional encodings once in log space.
        pe = torch.zeros(max_len, d_model)
        position = torch.arange(0, max_len).unsqueeze(1)
        div_term = torch.exp(torch.arange(0, d_model, 2) *
                             -(math.log(10000.0) / d_model))
        pe[:, 0::2] = torch.sin(position * div_term)
        pe[:, 1::2] = torch.cos(position * div_term)
        pe = pe.unsqueeze(0)
        self.register_buffer('pe', pe) 

def __init__(self, d_model, d_ff, dropout=0.1):
        super(PositionwiseFeedForward, self).__init__()
        self.w_1 = nn.Linear(d_model, d_ff)
        self.w_2 = nn.Linear(d_ff, d_model)
        self.dropout = nn.Dropout(dropout) 

def __init__(self, d_model, vocab):
        super(Generator, self).__init__()
        self.proj = nn.Linear(d_model, vocab) 

def clones(module, N):
    """Produce N identical layers."""
    return nn.ModuleList([copy.deepcopy(module) for _ in range(N)]) 

def __init__(self, latent_size, output_size):
        super().__init__()
        self.fc1 = nn.Linear(latent_size, 100)
        self.relu1 = nn.LeakyReLU(0.2, )
        self.fc2 = nn.Linear(100, 50)
        self.relu2 = nn.LeakyReLU(0.2)
        self.fc3 = nn.Linear(50, output_size)
        self.sigmoid = nn.Sigmoid() 

def __init__(self, encoder, decoder, src_embed, tgt_embed, generator, position_layer, model_size, latent_size):
        super(EncoderDecoder, self).__init__()
        self.encoder = encoder
        self.decoder = decoder
        self.src_embed = src_embed
        self.tgt_embed = tgt_embed
        self.generator = generator
        self.position_layer = position_layer
        self.model_size = model_size
        self.latent_size = latent_size
        self.sigmoid = nn.Sigmoid()

        # self.memory2latent = nn.Linear(self.model_size, self.latent_size)
        # self.latent2memory = nn.Linear(self.latent_size, self.model_size) 

def __init__(self, features, eps=1e-6):
        super(LayerNorm, self).__init__()
        self.a_2 = nn.Parameter(torch.ones(features))
        self.b_2 = nn.Parameter(torch.zeros(features))
        self.eps = eps 

def __init__(self, d_model, vocab):
        super(Embeddings, self).__init__()
        self.lut = nn.Embedding(vocab, d_model)
        self.d_model = d_model 

def __init__(self, size, dropout):
        super(SublayerConnection, self).__init__()
        self.norm = LayerNorm(size)
        self.dropout = nn.Dropout(dropout) 

def __init__(self, features, eps=1e-6):
        super(LayerNorm, self).__init__()
        self.a_2 = nn.Parameter(torch.ones(features))
        self.b_2 = nn.Parameter(torch.zeros(features))
        self.eps = eps 

def __init__(self, encoder, decoder, gru, src_embed, tgt_embed, generator, input_size):
        super(EncoderDecoder, self).__init__()
        self.encoder = encoder
        self.decoder = decoder
        self.gru = gru
        self.src_embed = src_embed
        self.tgt_embed = tgt_embed
        self.generator = generator
        self.linear = nn.Linear(input_size * 2, input_size)
        self.attention = AttentionScore(input_size, input_size)
        self.gru_decoder = nn.GRU(input_size * 2, input_size, 1) 

def __init__(self, d_model, d_ff, dropout=0.1):
        super(PositionwiseFeedForward, self).__init__()
        self.w_1 = nn.Linear(d_model, d_ff)
        self.w_2 = nn.Linear(d_ff, d_model)
        self.dropout = nn.Dropout(dropout) 

def __init__(self, h, d_model, dropout=0.1):
        """Take in model size and number of heads."""
        super(MultiHeadedAttention, self).__init__()
        assert d_model % h == 0
        # We assume d_v always equals d_k
        self.d_k = d_model // h
        self.h = h
        self.linears = clones(nn.Linear(d_model, d_model), 4)
        self.attn = None
        self.dropout = nn.Dropout(p=dropout) 

def clones(module, N):
    """Produce N identical layers."""
    return nn.ModuleList([copy.deepcopy(module) for _ in range(N)]) 

def __init__(self, latent_size, output_size):
        super().__init__()
        self.fc1 = nn.Linear(latent_size, 100)
        self.relu1 = nn.LeakyReLU(0.2, )
        self.fc2 = nn.Linear(100, 50)
        self.relu2 = nn.LeakyReLU(0.2)
        self.fc3 = nn.Linear(50, output_size)
        self.sigmoid = nn.Sigmoid()