Contrastive Models

hivegraph contains implementations of several contrastive models for graph representation learning. Available methods are:-

• Deep Graph Contrastive Representation Learning (GRACE)
  • GRACE
    • GRACE.train_step()
    • GRACE.semi_loss()
    • GRACE.batched_semi_loss()
    • GRACE.loss()
    • GRACE.forward()
    • GRACE.compute_cosine_sim()
    • GRACE.project()
    • GRACE.normalize_with_temp()
    • GRACE.reset_parameters()