(encoder_model, dataloader)
| 137 | |
| 138 | |
| 139 | def test(encoder_model, dataloader): |
| 140 | encoder_model.eval() |
| 141 | x = [] |
| 142 | y = [] |
| 143 | for data in dataloader: |
| 144 | data = data.to('cuda') |
| 145 | if data.x is None: |
| 146 | num_nodes = data.batch.size(0) |
| 147 | data.x = torch.ones((num_nodes, 1), dtype=torch.float32, device=data.batch.device) |
| 148 | g1, g2, _, _, _, _ = encoder_model(data.x, data.edge_index, batch=data.batch) |
| 149 | z = torch.cat([g1, g2], dim=1) |
| 150 | x.append(z) |
| 151 | y.append(data.y) |
| 152 | x = torch.cat(x, dim=0) |
| 153 | y = torch.cat(y, dim=0) |
| 154 | |
| 155 | split = get_split(num_samples=x.size()[0], train_ratio=0.8, test_ratio=0.1) |
| 156 | result = SVMEvaluator(linear=True)(x, y, split) |
| 157 | return result |
| 158 | |
| 159 | |
| 160 | def main(): |
no test coverage detected