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):
        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 _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 __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 _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 __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 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_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_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(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 32, 3, 1)
        self.conv2 = nn.Conv2d(32, 64, 3, 1)
        self.fc1 = nn.Linear(9216, 128) 

def train(model_q, model_k, device, train_loader, queue, optimizer, epoch, temp=0.07):
    model_q.train()
    total_loss = 0

    for batch_idx, (data, target) in enumerate(train_loader):
        x_q = data[0]
        x_k = data[1]

        x_q, x_k = x_q.to(device), x_k.to(device)
        q = model_q(x_q)
        k = model_k(x_k)
        k = k.detach()

        N = data[0].shape[0]
        K = queue.shape[0]
        l_pos = torch.bmm(q.view(N,1,-1), k.view(N,-1,1))
        l_neg = torch.mm(q.view(N,-1), queue.T.view(-1,K))

        logits = torch.cat([l_pos.view(N, 1), l_neg], dim=1)

        labels = torch.zeros(N, dtype=torch.long)
        labels = labels.to(device)

        cross_entropy_loss = nn.CrossEntropyLoss()
        loss = cross_entropy_loss(logits/temp, labels)

        optimizer.zero_grad()
        loss.backward()
        optimizer.step()

        total_loss += loss.item()

        momentum_update(model_q, model_k)

        queue = queue_data(queue, k)
        queue = dequeue_data(queue)

    total_loss /= len(train_loader.dataset)

    print('Train Epoch: {} \tLoss: {:.6f}'.format(epoch, total_loss)) 

def conv3x3(in_planes, out_planes, stride=1, dilation=1):
    "3x3 convolution with padding"
    # here with dilation
    return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=1+(dilation-1)*(3-1), dilation=dilation, bias=False) 

def __init__(self, inplanes, planes, stride=1, dilation=1, downsample=None):
        super(BasicBlock, self).__init__()
        self.conv1 = conv3x3(inplanes, planes, stride)
        self.bn1 = nn.BatchNorm2d(planes)
        self.relu = nn.ReLU(inplace=True)
        self.conv2 = conv3x3(planes, planes)
        self.bn2 = nn.BatchNorm2d(planes)
        self.downsample = downsample
        self.stride = stride 

def __init__(self, hidden_size, eps=1e-12):
            """Construct a layernorm module in the TF style (epsilon inside the square root).
            """
            super(BertLayerNorm, self).__init__()
            self.weight = nn.Parameter(torch.ones(hidden_size))
            self.bias = nn.Parameter(torch.zeros(hidden_size))
            self.variance_epsilon = eps 

def __init__(self, config):
        super(BertEmbeddings, self).__init__()
        self.word_embeddings = nn.Embedding(config.vocab_size, config.hidden_size)
        self.position_embeddings = nn.Embedding(config.max_position_embeddings, config.hidden_size)
        self.token_type_embeddings = nn.Embedding(config.type_vocab_size, config.hidden_size)

        # self.LayerNorm is not snake-cased to stick with TensorFlow model variable name and be able to load
        # any TensorFlow checkpoint file
        self.LayerNorm = BertLayerNorm(config.hidden_size, eps=1e-12)
        self.dropout = nn.Dropout(config.hidden_dropout_prob) 

def __init__(self, config):
        super(BertSelfAttention, self).__init__()
        if config.hidden_size % config.num_attention_heads != 0:
            raise ValueError(
                "The hidden size (%d) is not a multiple of the number of attention "
                "heads (%d)" % (config.hidden_size, config.num_attention_heads))
        self.num_attention_heads = config.num_attention_heads
        self.attention_head_size = int(config.hidden_size / config.num_attention_heads)
        self.all_head_size = self.num_attention_heads * self.attention_head_size

        self.query = nn.Linear(config.hidden_size, self.all_head_size)
        self.key = nn.Linear(config.hidden_size, self.all_head_size)
        self.value = nn.Linear(config.hidden_size, self.all_head_size)

        self.dropout = nn.Dropout(config.attention_probs_dropout_prob) 

def forward(self, hidden_states, attention_mask):
        mixed_query_layer = self.query(hidden_states)
        mixed_key_layer = self.key(hidden_states)
        mixed_value_layer = self.value(hidden_states)

        query_layer = self.transpose_for_scores(mixed_query_layer)
        key_layer = self.transpose_for_scores(mixed_key_layer)
        value_layer = self.transpose_for_scores(mixed_value_layer)

        # Take the dot product between "query" and "key" to get the raw attention scores.
        attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2))
        attention_scores = attention_scores / math.sqrt(self.attention_head_size)
        # Apply the attention mask is (precomputed for all layers in BertModel forward() function)
        attention_scores = attention_scores + attention_mask

        # Normalize the attention scores to probabilities.
        attention_probs = nn.Softmax(dim=-1)(attention_scores)

        # This is actually dropping out entire tokens to attend to, which might
        # seem a bit unusual, but is taken from the original Transformer paper.
        attention_probs = self.dropout(attention_probs)

        context_layer = torch.matmul(attention_probs, value_layer)
        context_layer = context_layer.permute(0, 2, 1, 3).contiguous()
        new_context_layer_shape = context_layer.size()[:-2] + (self.all_head_size,)
        context_layer = context_layer.view(*new_context_layer_shape)
        return context_layer 

def __init__(self, config):
        super(BertSelfOutput, self).__init__()
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        self.LayerNorm = BertLayerNorm(config.hidden_size, eps=1e-12)
        self.dropout = nn.Dropout(config.hidden_dropout_prob) 

def __init__(self, config):
        super(BertOutput, self).__init__()
        self.dense = nn.Linear(config.intermediate_size, config.hidden_size)
        self.LayerNorm = BertLayerNorm(config.hidden_size, eps=1e-12)
        self.dropout = nn.Dropout(config.hidden_dropout_prob) 

def __init__(self, config):
        super(BertEncoder, self).__init__()
        layer = BertLayer(config)
        self.layer = nn.ModuleList([copy.deepcopy(layer) for _ in range(config.num_hidden_layers)]) 

def __init__(self, config):
        super(BertPooler, self).__init__()
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        self.activation = nn.Tanh() 

def __init__(self, config):
        super(BertPredictionHeadTransform, self).__init__()
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        self.transform_act_fn = ACT2FN[config.hidden_act] \
            if isinstance(config.hidden_act, str) else config.hidden_act
        self.LayerNorm = BertLayerNorm(config.hidden_size, eps=1e-12) 

def __init__(self, config, bert_model_embedding_weights):
        super(BertLMPredictionHead, self).__init__()
        self.transform = BertPredictionHeadTransform(config)

        # The output weights are the same as the input embeddings, but there is
        # an output-only bias for each token.
        self.decoder = nn.Linear(bert_model_embedding_weights.size(1),
                                 bert_model_embedding_weights.size(0),
                                 bias=False)
        self.decoder.weight = bert_model_embedding_weights
        self.bias = nn.Parameter(torch.zeros(bert_model_embedding_weights.size(0))) 

def __init__(self, config, bert_model_embedding_weights):
        super(BertPreTrainingHeads, self).__init__()
        self.predictions = BertLMPredictionHead(config, bert_model_embedding_weights)
        self.seq_relationship = nn.Linear(config.hidden_size, 2) 

def init_bert_weights(self, module):
        """ Initialize the weights.
        """
        if isinstance(module, (nn.Linear, nn.Embedding)):
            # Slightly different from the TF version which uses truncated_normal for initialization
            # cf https://github.com/pytorch/pytorch/pull/5617
            module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
        elif isinstance(module, BertLayerNorm):
            module.bias.data.normal_(mean=0.0, std=self.config.initializer_range)
            module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
        if isinstance(module, nn.Linear) and module.bias is not None:
            module.bias.data.zero_() 

def __init__(self, config, num_labels=2):
        super(BertForSequenceClassification, self).__init__(config)
        self.num_labels = num_labels
        self.bert = BertModel(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.classifier = nn.Linear(config.hidden_size, num_labels)
        self.apply(self.init_bert_weights) 

def __init__(self, config, num_choices=2):
        super(BertForMultipleChoice, self).__init__(config)
        self.num_choices = num_choices
        self.bert = BertModel(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.classifier = nn.Linear(config.hidden_size, 1)
        self.apply(self.init_bert_weights) 

def __init__(self, config, num_labels=2):
        super(BertForTokenClassification, self).__init__(config)
        self.num_labels = num_labels
        self.bert = BertModel(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.classifier = nn.Linear(config.hidden_size, num_labels)
        self.apply(self.init_bert_weights) 

def __init__(self, config):
        super(BertForQuestionAnswering, self).__init__(config)
        self.bert = BertModel(config)
        self.qa_outputs = nn.Linear(config.hidden_size, 1)
        self.apply(self.init_bert_weights)