DropEdge¶

class dgl.transforms.DropEdge(p=0.5)[source]¶

Bases: dgl.transforms.module.BaseTransform

Randomly drop edges, as described in DropEdge: Towards Deep Graph Convolutional Networks on Node Classification and Graph Contrastive Learning with Augmentations.

Parameters: p (float, optional) – Probability of an edge to be dropped.

Example

>>> import dgl
>>> import torch
>>> from dgl import DropEdge

>>> transform = DropEdge()
>>> g = dgl.rand_graph(5, 20)
>>> g.edata['h'] = torch.arange(g.num_edges())
>>> new_g = transform(g)
>>> print(new_g)
Graph(num_nodes=5, num_edges=12,
      ndata_schemes={}
      edata_schemes={'h': Scheme(shape=(), dtype=torch.int64)})
>>> print(new_g.edata['h'])
tensor([0, 1, 3, 7, 8, 10, 11, 12, 13, 15, 18, 19])