[NN][Model] GATv2 (#3473)

* [Model][Core] GATv2

* lint

* gatv2conv.py

* lint

* lint

* style and docs

* lint

* gatv2conv fix
Co-authored-by: Shaked Brody shakedbr@campus.technion.ac.il <shakedbr@tangerine.cslcs.technion.ac.il>
Co-authored-by: Mufei Li <mufeili1996@gmail.com>

README.md

@@ -288,6 +288,8 @@ Take the survey [here](https://forms.gle/Ej3jHCocACmb49Gp8) and leave any feedba
 
 1. [**GNNLens: A Visual Analytics Approach for Prediction Error Diagnosis of Graph Neural Networks**](https://arxiv.org/abs/2011.11048v5), *Zhihua Jin, Yong Wang, Qianwen Wang, Yao Ming, Tengfei Ma, Huamin Qu*
 
+1. [**How Attentive are Graph Attention Networks?**](https://arxiv.org/pdf/2105.14491.pdf), *Shaked Brody, Uri Alon, Eran Yahav*, [code](https://github.com/tech-srl/how_attentive_are_gats)
+
 </details>
 
 ## Contributing

docs/source/api/python/nn.pytorch.rst

@@ -45,6 +45,12 @@ GATConv
     :members: forward
     :show-inheritance:
     
+GATv2Conv
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. autoclass:: dgl.nn.pytorch.conv.GATv2Conv
+    :members: forward
+    :show-inheritance:
 
 EGATConv
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

examples/README.md

@@ -23,6 +23,9 @@ To quickly locate the examples of your interest, search for the tagged keywords
 - <a name="pct"></a> Guo et al. PCT: Point cloud transformer. [Paper link](http://arxiv.org/abs/2012.09688).
     - Example code: [PyTorch](../examples/pytorch/pointcloud/pct)
     - Tags: point cloud classification, point cloud part-segmentation
+- <a name='gatv2'></a> Brody et al. How Attentive are Graph Attention Networks? [Paper link](https://arxiv.org/abs/2105.14491).
+    - Example code: [PyTorch](../examples/pytorch/gatv2)
+    - Tags: graph attention, gat, gatv2, attention
 
 ## 2020
 - <a name="eeg-gcnn"></a> Wagh et al. EEG-GCNN: Augmenting Electroencephalogram-based Neurological Disease Diagnosis using a Domain-guided Graph Convolutional Neural Network. [Paper link](http://proceedings.mlr.press/v136/wagh20a.html). 

examples/pytorch/gatv2/README.md

+Graph Attention Networks v2 (GATv2)
+============
+
+- Paper link: [How Attentive are Graph Attention Networks?](https://arxiv.org/pdf/2105.14491.pdf)
+- Author's code repo: [https://github.com/tech-srl/how_attentive_are_gats](https://github.com/tech-srl/how_attentive_are_gats).
+- Annotated implemetnation: [https://nn.labml.ai/graphs/gatv2/index.html]
+
+Dependencies
+------------
+- torch
+- requests
+- sklearn
+
+How to run
+----------
+
+Run with following:
+
+```bash
+python3 train.py --dataset=cora
+```
+
+```bash
+python3 train.py --dataset=citeseer
+```
+
+```bash
+python3 train.py --dataset=pubmed
+```
+
+Results
+-------
+
+| Dataset  | Test Accuracy |
+| -------- | ------------- |
+| Cora     |  82.10        |
+| Citeseer |  70.00        |
+| Pubmed   |  77.2         |
+
+* All the accuracy numbers are obtained after 200 epochs.
\ No newline at end of file

examples/pytorch/gatv2/gatv2.py

+"""
+Graph Attention Networks in DGL using SPMV optimization.
+References
+----------
+Paper: https://arxiv.org/pdf/2105.14491.pdf
+Author's code: https://github.com/tech-srl/how_attentive_are_gats
+"""
+
+import torch
+import torch.nn as nn
+from dgl.nn import GATv2Conv
+    
+
+class GATv2(nn.Module):
+    def __init__(self,
+                 num_layers,
+                 in_dim,
+                 num_hidden,
+                 num_classes,
+                 heads,
+                 activation,
+                 feat_drop,
+                 attn_drop,
+                 negative_slope,
+                 residual):
+        super(GATv2, self).__init__()
+        self.num_layers = num_layers
+        self.gatv2_layers = nn.ModuleList()
+        self.activation = activation
+        # input projection (no residual)
+        self.gatv2_layers.append(GATv2Conv(
+            in_dim, num_hidden, heads[0],
+            feat_drop, attn_drop, negative_slope, False, self.activation, bias=False, share_weights=True))
+        # hidden layers
+        for l in range(1, num_layers):
+            # due to multi-head, the in_dim = num_hidden * num_heads
+            self.gatv2_layers.append(GATv2Conv(
+                num_hidden * heads[l-1], num_hidden, heads[l],
+                feat_drop, attn_drop, negative_slope, residual, self.activation, bias=False, share_weights=True))
+        # output projection
+        self.gatv2_layers.append(GATv2Conv(
+            num_hidden * heads[-2], num_classes, heads[-1],
+            feat_drop, attn_drop, negative_slope, residual, None, bias=False, share_weights=True))
+
+    def forward(self, g, inputs):
+        h = inputs
+        for l in range(self.num_layers):
+            h = self.gatv2_layers[l](h).flatten(1)
+        # output projection
+        logits = self.gatv2_layers[-1](h).mean(1)
+        return logits

examples/pytorch/gatv2/train.py

+"""
+Graph Attention Networks v2 (GATv2) in DGL using SPMV optimization.
+Multiple heads are also batched together for faster training.
+"""
+
+import argparse
+import numpy as np
+import time
+import torch
+import torch.nn.functional as F
+import dgl
+from dgl.data import register_data_args
+from dgl.data import CoraGraphDataset, CiteseerGraphDataset, PubmedGraphDataset
+
+from gatv2 import GATv2
+
+
+class EarlyStopping:
+    def __init__(self, patience=10):
+        self.patience = patience
+        self.counter = 0
+        self.best_score = None
+        self.early_stop = False
+
+    def step(self, acc, model):
+        score = acc
+        if self.best_score is None:
+            self.best_score = score
+            self.save_checkpoint(model)
+        elif score < self.best_score:
+            self.counter += 1
+            print(f'EarlyStopping counter: {self.counter} out of {self.patience}')
+            if self.counter >= self.patience:
+                self.early_stop = True
+        else:
+            self.best_score = score
+            self.save_checkpoint(model)
+            self.counter = 0
+        return self.early_stop
+
+    def save_checkpoint(self, model):
+        '''Saves model when validation loss decrease.'''
+        torch.save(model.state_dict(), 'es_checkpoint.pt')
+
+def accuracy(logits, labels):
+    _, indices = torch.max(logits, dim=1)
+    correct = torch.sum(indices == labels)
+    return correct.item() * 1.0 / len(labels)
+
+
+def evaluate(model, g, features, labels, mask):
+    model.eval()
+    with torch.no_grad():
+        logits = model(g, features)
+        logits = logits[mask]
+        labels = labels[mask]
+        return accuracy(logits, labels)
+
+
+def main(args):
+    # load and preprocess dataset
+    if args.dataset == 'cora':
+        data = CoraGraphDataset()
+    elif args.dataset == 'citeseer':
+        data = CiteseerGraphDataset()
+    elif args.dataset == 'pubmed':
+        data = PubmedGraphDataset()
+    else:
+        raise ValueError('Unknown dataset: {}'.format(args.dataset))
+
+    g = data[0]
+    if args.gpu < 0:
+        cuda = False
+    else:
+        cuda = True
+        g = g.int().to(args.gpu)
+
+    features = g.ndata['feat']
+    labels = g.ndata['label']
+    train_mask = g.ndata['train_mask']
+    val_mask = g.ndata['val_mask']
+    test_mask = g.ndata['test_mask']
+    num_feats = features.shape[1]
+    n_classes = data.num_labels
+    n_edges = data.graph.number_of_edges()
+    print("""----Data statistics------'
+      #Edges %d
+      #Classes %d
+      #Train samples %d
+      #Val samples %d
+      #Test samples %d""" %
+          (n_edges, n_classes,
+           train_mask.int().sum().item(),
+           val_mask.int().sum().item(),
+           test_mask.int().sum().item()))
+
+    # add self loop
+    g = dgl.remove_self_loop(g)
+    g = dgl.add_self_loop(g)
+    n_edges = g.number_of_edges()
+    # create model
+    heads = ([args.num_heads] * args.num_layers) + [args.num_out_heads]
+    model = GATv2(args.num_layers,
+                  num_feats,
+                  args.num_hidden,
+                  n_classes,
+                  heads,
+                  F.elu,
+                  args.in_drop,
+                  args.attn_drop,
+                  args.negative_slope,
+                  args.residual)
+    print(model)
+    if args.early_stop:
+        stopper = EarlyStopping(patience=100)
+    if cuda:
+        model.cuda()
+    loss_fcn = torch.nn.CrossEntropyLoss()
+
+    # use optimizer
+    optimizer = torch.optim.Adam(
+        model.parameters(), lr=args.lr, weight_decay=args.weight_decay)
+
+    # initialize graph
+    dur = []
+    for epoch in range(args.epochs):
+        model.train()
+        if epoch >= 3:
+            t0 = time.time()
+        # forward
+        logits = model(g, features)
+        loss = loss_fcn(logits[train_mask], labels[train_mask])
+
+        optimizer.zero_grad()
+        loss.backward()
+        optimizer.step()
+
+        if epoch >= 3:
+            dur.append(time.time() - t0)
+
+        train_acc = accuracy(logits[train_mask], labels[train_mask])
+
+        if args.fastmode:
+            val_acc = accuracy(logits[val_mask], labels[val_mask])
+        else:
+            val_acc = evaluate(g, model, features, labels, val_mask)
+            if args.early_stop:
+                if stopper.step(val_acc, model):
+                    break
+
+        print("Epoch {:05d} | Time(s) {:.4f} | Loss {:.4f} | TrainAcc {:.4f} |"
+              " ValAcc {:.4f} | ETputs(KTEPS) {:.2f}".
+              format(epoch, np.mean(dur), loss.item(), train_acc,
+                     val_acc, n_edges / np.mean(dur) / 1000))
+
+    print()
+    if args.early_stop:
+        model.load_state_dict(torch.load('es_checkpoint.pt'))
+    acc = evaluate(model, features, labels, test_mask)
+    print("Test Accuracy {:.4f}".format(acc))
+
+
+if __name__ == '__main__':
+
+    parser = argparse.ArgumentParser(description='GAT')
+    register_data_args(parser)
+    parser.add_argument("--gpu", type=int, default=-1,
+                        help="which GPU to use. Set -1 to use CPU.")
+    parser.add_argument("--epochs", type=int, default=200,
+                        help="number of training epochs")
+    parser.add_argument("--num-heads", type=int, default=8,
+                        help="number of hidden attention heads")
+    parser.add_argument("--num-out-heads", type=int, default=1,
+                        help="number of output attention heads")
+    parser.add_argument("--num-layers", type=int, default=1,
+                        help="number of hidden layers")
+    parser.add_argument("--num-hidden", type=int, default=8,
+                        help="number of hidden units")
+    parser.add_argument("--residual", action="store_true", default=False,
+                        help="use residual connection")
+    parser.add_argument("--in-drop", type=float, default=.7,
+                        help="input feature dropout")
+    parser.add_argument("--attn-drop", type=float, default=.7,
+                        help="attention dropout")
+    parser.add_argument("--lr", type=float, default=0.005,
+                        help="learning rate")
+    parser.add_argument('--weight-decay', type=float, default=5e-4,
+                        help="weight decay")
+    parser.add_argument('--negative-slope', type=float, default=0.2,
+                        help="the negative slope of leaky relu")
+    parser.add_argument('--early-stop', action='store_true', default=False,
+                        help="indicates whether to use early stop or not")
+    parser.add_argument('--fastmode', action="store_true", default=False,
+                        help="skip re-evaluate the validation set")
+    args = parser.parse_args()
+    print(args)
+
+    main(args)

python/dgl/nn/pytorch/conv/__init__.py

@@ -6,6 +6,7 @@ from .appnpconv import APPNPConv
 from .chebconv import ChebConv
 from .edgeconv import EdgeConv
 from .gatconv import GATConv
+from .gatv2conv import GATv2Conv
 from .egatconv import EGATConv
 from .ginconv import GINConv
 from .gmmconv import GMMConv
@@ -25,8 +26,8 @@ from .dotgatconv import DotGatConv
 from .twirlsconv import TWIRLSConv, TWIRLSUnfoldingAndAttention
 from .gcn2conv import GCN2Conv
 
-__all__ = ['GraphConv', 'EdgeWeightNorm', 'GATConv', 'EGATConv', 'TAGConv', 'RelGraphConv',
-           'SAGEConv', 'SGConv', 'APPNPConv', 'GINConv', 'GatedGraphConv', 'GMMConv',
-           'ChebConv', 'AGNNConv', 'NNConv', 'DenseGraphConv', 'DenseSAGEConv',
+__all__ = ['GraphConv', 'EdgeWeightNorm', 'GATConv', 'GATv2Conv', 'EGATConv', 'TAGConv',
+           'RelGraphConv', 'SAGEConv', 'SGConv', 'APPNPConv', 'GINConv', 'GatedGraphConv',
+           'GMMConv', 'ChebConv', 'AGNNConv', 'NNConv', 'DenseGraphConv', 'DenseSAGEConv',
            'DenseChebConv', 'EdgeConv', 'AtomicConv', 'CFConv', 'DotGatConv', 'TWIRLSConv',
            'TWIRLSUnfoldingAndAttention', 'GCN2Conv']

python/dgl/nn/pytorch/conv/gatv2conv.py

+"""Torch modules for graph attention networks v2 (GATv2)."""
+# pylint: disable= no-member, arguments-differ, invalid-name
+import torch as th
+from torch import nn
+
+from .... import function as fn
+from ...functional import edge_softmax
+from ....base import DGLError
+from ..utils import Identity
+from ....utils import expand_as_pair
+
+# pylint: enable=W0235
+class GATv2Conv(nn.Module):
+    r"""
+
+    Description
+    -----------
+    Apply GATv2 from
+    `How Attentive are Graph Attention Networks? <https://arxiv.org/pdf/2105.14491.pdf>`__
+    over an input signal.
+
+    .. math::
+        h_i^{(l+1)} = \sum_{j\in \mathcal{N}(i)} \alpha_{i,j} W^{(l)}_{right} h_j^{(l)}
+
+    where :math:`\alpha_{ij}` is the attention score bewteen node :math:`i` and
+    node :math:`j`:
+
+    .. math::
+        \alpha_{ij}^{l} &= \mathrm{softmax_i} (e_{ij}^{l})
+
+        e_{ij}^{l} &= \vec{a}^T\mathrm{LeakyReLU}\left(
+            W^{(l)}_{left} h_{i} + W^{(l)}_{right} h_{j}]\right)
+
+    Parameters
+    ----------
+    in_feats : int, or pair of ints
+        Input feature size; i.e, the number of dimensions of :math:`h_i^{(l)}`.
+        If the layer is to be applied to a unidirectional bipartite graph, `in_feats`
+        specifies the input feature size on both the source and destination nodes.
+        If a scalar is given, the source and destination node feature size
+        would take the same value.
+    out_feats : int
+        Output feature size; i.e, the number of dimensions of :math:`h_i^{(l+1)}`.
+    num_heads : int
+        Number of heads in Multi-Head Attention.
+    feat_drop : float, optional
+        Dropout rate on feature. Defaults: ``0``.
+    attn_drop : float, optional
+        Dropout rate on attention weight. Defaults: ``0``.
+    negative_slope : float, optional
+        LeakyReLU angle of negative slope. Defaults: ``0.2``.
+    residual : bool, optional
+        If True, use residual connection. Defaults: ``False``.
+    activation : callable activation function/layer or None, optional.
+        If not None, applies an activation function to the updated node features.
+        Default: ``None``.
+    allow_zero_in_degree : bool, optional
+        If there are 0-in-degree nodes in the graph, output for those nodes will be invalid
+        since no message will be passed to those nodes. This is harmful for some applications
+        causing silent performance regression. This module will raise a DGLError if it detects
+        0-in-degree nodes in input graph. By setting ``True``, it will suppress the check
+        and let the users handle it by themselves. Defaults: ``False``.
+    bias : bool, optional
+        If set to :obj:`False`, the layer will not learn
+        an additive bias. (default: :obj:`True`)
+    share_weights : bool, optional
+        If set to :obj:`True`, the same matrix for :math:`W_{left}` and :math:`W_{right}` in
+        the above equations, will be applied to the source and the target node of every edge.
+        (default: :obj:`False`)
+
+    Note
+    ----
+    Zero in-degree nodes will lead to invalid output value. This is because no message
+    will be passed to those nodes, the aggregation function will be applied on empty input.
+    A common practice to avoid this is to add a self-loop for each node in the graph if
+    it is homogeneous, which can be achieved by:
+
+    >>> g = ... # a DGLGraph
+    >>> g = dgl.add_self_loop(g)
+
+    Calling ``add_self_loop`` will not work for some graphs, for example, heterogeneous graph
+    since the edge type can not be decided for self_loop edges. Set ``allow_zero_in_degree``
+    to ``True`` for those cases to unblock the code and handle zero-in-degree nodes manually.
+    A common practise to handle this is to filter out the nodes with zero-in-degree when use
+    after conv.
+
+    Examples
+    --------
+    >>> import dgl
+    >>> import numpy as np
+    >>> import torch as th
+    >>> from dgl.nn import GATv2Conv
+
+    >>> # Case 1: Homogeneous graph
+    >>> g = dgl.graph(([0,1,2,3,2,5], [1,2,3,4,0,3]))
+    >>> g = dgl.add_self_loop(g)
+    >>> feat = th.ones(6, 10)
+    >>> gatv2conv = GATv2Conv(10, 2, num_heads=3)
+    >>> res = gatv2conv(g, feat)
+    >>> res
+    tensor([[[ 1.9599,  1.0239],
+            [ 3.2015, -0.5512],
+            [ 2.3700, -2.2182]],
+            [[ 1.9599,  1.0239],
+            [ 3.2015, -0.5512],
+            [ 2.3700, -2.2182]],
+            [[ 1.9599,  1.0239],
+            [ 3.2015, -0.5512],
+            [ 2.3700, -2.2182]],
+            [[ 1.9599,  1.0239],
+            [ 3.2015, -0.5512],
+            [ 2.3700, -2.2182]],
+            [[ 1.9599,  1.0239],
+            [ 3.2015, -0.5512],
+            [ 2.3700, -2.2182]],
+            [[ 1.9599,  1.0239],
+            [ 3.2015, -0.5512],
+            [ 2.3700, -2.2182]]], grad_fn=<GSpMMBackward>)
+
+    >>> # Case 2: Unidirectional bipartite graph
+    >>> u = [0, 1, 0, 0, 1]
+    >>> v = [0, 1, 2, 3, 2]
+    >>> g = dgl.heterograph({('A', 'r', 'B'): (u, v)})
+    >>> u_feat = th.tensor(np.random.rand(2, 5).astype(np.float32))
+    >>> v_feat = th.tensor(np.random.rand(4, 10).astype(np.float32))
+    >>> gatv2conv = GATv2Conv((5,10), 2, 3)
+    >>> res = gatv2conv(g, (u_feat, v_feat))
+    >>> res
+    tensor([[[-0.0935, -0.4273],
+            [-1.1850,  0.1123],
+            [-0.2002,  0.1155]],
+            [[ 0.1908, -1.2095],
+            [-0.0129,  0.6408],
+            [-0.8135,  0.1157]],
+            [[ 0.0596, -0.8487],
+            [-0.5421,  0.4022],
+            [-0.4805,  0.1156]],
+            [[-0.0935, -0.4273],
+            [-1.1850,  0.1123],
+            [-0.2002,  0.1155]]], grad_fn=<GSpMMBackward>)
+    """
+    def __init__(self,
+                 in_feats,
+                 out_feats,
+                 num_heads,
+                 feat_drop=0.,
+                 attn_drop=0.,
+                 negative_slope=0.2,
+                 residual=False,
+                 activation=None,
+                 allow_zero_in_degree=False,
+                 bias=True,
+                 share_weights=False):
+        super(GATv2Conv, self).__init__()
+        self._num_heads = num_heads
+        self._in_src_feats, self._in_dst_feats = expand_as_pair(in_feats)
+        self._out_feats = out_feats
+        self._allow_zero_in_degree = allow_zero_in_degree
+        if isinstance(in_feats, tuple):
+            self.fc_src = nn.Linear(
+                self._in_src_feats, out_feats * num_heads, bias=bias)
+            self.fc_dst = nn.Linear(
+                self._in_dst_feats, out_feats * num_heads, bias=bias)
+        else:
+            self.fc_src = nn.Linear(
+                self._in_src_feats, out_feats * num_heads, bias=bias)
+            if share_weights:
+                self.fc_dst = self.fc_src
+            else:
+                self.fc_dst = nn.Linear(
+                    self._in_src_feats, out_feats * num_heads, bias=bias)
+        self.attn = nn.Parameter(th.FloatTensor(size=(1, num_heads, out_feats)))
+        self.feat_drop = nn.Dropout(feat_drop)
+        self.attn_drop = nn.Dropout(attn_drop)
+        self.leaky_relu = nn.LeakyReLU(negative_slope)
+        if residual:
+            if self._in_dst_feats != out_feats:
+                self.res_fc = nn.Linear(
+                    self._in_dst_feats, num_heads * out_feats, bias=bias)
+            else:
+                self.res_fc = Identity()
+        else:
+            self.register_buffer('res_fc', None)
+        self.activation = activation
+        self.share_weights = share_weights
+        self.bias = bias
+        self.reset_parameters()
+
+    def reset_parameters(self):
+        """
+        Description
+        -----------
+        Reinitialize learnable parameters.
+
+        Note
+        ----
+        The fc weights :math:`W^{(l)}` are initialized using Glorot uniform initialization.
+        The attention weights are using xavier initialization method.
+        """
+        gain = nn.init.calculate_gain('relu')
+        nn.init.xavier_normal_(self.fc_src.weight, gain=gain)
+        if self.bias:
+            nn.init.constant_(self.fc_src.bias, 0)
+        if not self.share_weights:
+            nn.init.xavier_normal_(self.fc_dst.weight, gain=gain)
+            if self.bias:
+                nn.init.constant_(self.fc_dst.bias, 0)
+        nn.init.xavier_normal_(self.attn, gain=gain)
+        if isinstance(self.res_fc, nn.Linear):
+            nn.init.xavier_normal_(self.res_fc.weight, gain=gain)
+            if self.bias:
+                nn.init.constant_(self.res_fc.bias, 0)
+
+    def set_allow_zero_in_degree(self, set_value):
+        r"""
+        Description
+        -----------
+        Set allow_zero_in_degree flag.
+
+        Parameters
+        ----------
+        set_value : bool
+            The value to be set to the flag.
+        """
+        self._allow_zero_in_degree = set_value
+
+    def forward(self, graph, feat, get_attention=False):
+        r"""
+        Description
+        -----------
+        Compute graph attention network layer.
+
+        Parameters
+        ----------
+        graph : DGLGraph
+            The graph.
+        feat : torch.Tensor or pair of torch.Tensor
+            If a torch.Tensor is given, the input feature of shape :math:`(N, D_{in})` where
+            :math:`D_{in}` is size of input feature, :math:`N` is the number of nodes.
+            If a pair of torch.Tensor is given, the pair must contain two tensors of shape
+            :math:`(N_{in}, D_{in_{src}})` and :math:`(N_{out}, D_{in_{dst}})`.
+        get_attention : bool, optional
+            Whether to return the attention values. Default to False.
+
+        Returns
+        -------
+        torch.Tensor
+            The output feature of shape :math:`(N, H, D_{out})` where :math:`H`
+            is the number of heads, and :math:`D_{out}` is size of output feature.
+        torch.Tensor, optional
+            The attention values of shape :math:`(E, H, 1)`, where :math:`E` is the number of
+            edges. This is returned only when :attr:`get_attention` is ``True``.
+
+        Raises
+        ------
+        DGLError
+            If there are 0-in-degree nodes in the input graph, it will raise DGLError
+            since no message will be passed to those nodes. This will cause invalid output.
+            The error can be ignored by setting ``allow_zero_in_degree`` parameter to ``True``.
+        """
+        with graph.local_scope():
+            if not self._allow_zero_in_degree:
+                if (graph.in_degrees() == 0).any():
+                    raise DGLError('There are 0-in-degree nodes in the graph, '
+                                   'output for those nodes will be invalid. '
+                                   'This is harmful for some applications, '
+                                   'causing silent performance regression. '
+                                   'Adding self-loop on the input graph by '
+                                   'calling `g = dgl.add_self_loop(g)` will resolve '
+                                   'the issue. Setting ``allow_zero_in_degree`` '
+                                   'to be `True` when constructing this module will '
+                                   'suppress the check and let the code run.')
+
+            if isinstance(feat, tuple):
+                h_src = self.feat_drop(feat[0])
+                h_dst = self.feat_drop(feat[1])
+                feat_src = self.fc_src(h_src).view(-1, self._num_heads, self._out_feats)
+                feat_dst = self.fc_dst(h_dst).view(-1, self._num_heads, self._out_feats)
+            else:
+                h_src = h_dst = self.feat_drop(feat)
+                feat_src = self.fc_src(h_src).view(
+                    -1, self._num_heads, self._out_feats)
+                if self.share_weights:
+                    feat_dst = feat_src
+                else:
+                    feat_dst = self.fc_dst(h_src).view(
+                        -1, self._num_heads, self._out_feats)
+                if graph.is_block:
+                    feat_dst = feat_src[:graph.number_of_dst_nodes()]
+            graph.srcdata.update({'el': feat_src})# (num_src_edge, num_heads, out_dim)
+            graph.dstdata.update({'er': feat_dst})
+            graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
+            e = self.leaky_relu(graph.edata.pop('e'))# (num_src_edge, num_heads, out_dim)
+            e = (e * self.attn).sum(dim=-1).unsqueeze(dim=2)# (num_edge, num_heads, 1)
+            # compute softmax
+            graph.edata['a'] = self.attn_drop(edge_softmax(graph, e)) # (num_edge, num_heads)
+            # message passing
+            graph.update_all(fn.u_mul_e('el', 'a', 'm'),
+                             fn.sum('m', 'ft'))
+            rst = graph.dstdata['ft']
+            # residual
+            if self.res_fc is not None:
+                resval = self.res_fc(h_dst).view(h_dst.shape[0], -1, self._out_feats)
+                rst = rst + resval
+            # activation
+            if self.activation:
+                rst = self.activation(rst)
+
+            if get_attention:
+                return rst, graph.edata['a']
+            else:
+                return rst

tests/pytorch/test_nn.py

@@ -564,6 +564,45 @@ def test_gat_conv_bi(g, idtype, out_dim, num_heads):
     _, a = gat(g, feat, get_attention=True)
     assert a.shape == (g.number_of_edges(), num_heads, 1)
 
+@parametrize_dtype
+@pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
+@pytest.mark.parametrize('out_dim', [1, 5])
+@pytest.mark.parametrize('num_heads', [1, 4])
+def test_gatv2_conv(g, idtype, out_dim, num_heads):
+    g = g.astype(idtype).to(F.ctx())
+    ctx = F.ctx()
+    gat = nn.GATv2Conv(5, out_dim, num_heads)
+    feat = F.randn((g.number_of_src_nodes(), 5))
+    gat = gat.to(ctx)
+    h = gat(g, feat)
+
+    # test pickle
+    th.save(gat, tmp_buffer)
+
+    assert h.shape == (g.number_of_dst_nodes(), num_heads, out_dim)
+    _, a = gat(g, feat, get_attention=True)
+    assert a.shape == (g.number_of_edges(), num_heads, 1)
+
+    # test residual connection
+    gat = nn.GATConv(5, out_dim, num_heads, residual=True)
+    gat = gat.to(ctx)
+    h = gat(g, feat)
+
+@parametrize_dtype
+@pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
+@pytest.mark.parametrize('out_dim', [1, 2])
+@pytest.mark.parametrize('num_heads', [1, 4])
+def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
+    g = g.astype(idtype).to(F.ctx())
+    ctx = F.ctx()
+    gat = nn.GATv2Conv(5, out_dim, num_heads)
+    feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
+    gat = gat.to(ctx)
+    h = gat(g, feat)
+    assert h.shape == (g.number_of_dst_nodes(), num_heads, out_dim)
+    _, a = gat(g, feat, get_attention=True)
+    assert a.shape == (g.number_of_edges(), num_heads, 1)
+
 @parametrize_dtype
 @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
 @pytest.mark.parametrize('out_node_feats', [1, 5])
@@ -1159,6 +1198,7 @@ if __name__ == '__main__':
     test_rgcn_sorted()
     test_tagconv()
     test_gat_conv()
+    test_gatv2_conv()
     test_egat_conv()
     test_sage_conv()
     test_sgc_conv()