(self, x, y, **kwargs)
| 72 | super(Curriculum_SimpleModel, self).__init__(hidden_dim, empty_grad) |
| 73 | |
| 74 | def forward(self, x, y, **kwargs): |
| 75 | seqlen = kwargs.get('curriculum_seqlen', None) |
| 76 | loss = super(Curriculum_SimpleModel, self).forward(x, y) |
| 77 | return loss, seqlen |
| 78 | |
| 79 | |
| 80 | class SimpleMoEModel(torch.nn.Module): |