assume target and nerf rgb are inferenced at the same time, slice target batch and nerf batch and output stacked features :param x: image blob (2B x C x H x W) :return feature: (2 x B x C x H x W)
(self, x)
| 413 | return feature |
| 414 | |
| 415 | def _aggregate_feature2(self, x): |
| 416 | ''' |
| 417 | assume target and nerf rgb are inferenced at the same time, |
| 418 | slice target batch and nerf batch and output stacked features |
| 419 | :param x: image blob (2B x C x H x W) |
| 420 | :return feature: (2 x B x C x H x W) |
| 421 | ''' |
| 422 | batch = x.shape[0] # should be target batch_size + rgb batch_size |
| 423 | feature_t = x[:batch//2] |
| 424 | feature_r = x[batch//2:] |
| 425 | feature = torch.stack([feature_t, feature_r]) |
| 426 | return feature |
| 427 | |
| 428 | def forward(self, x, upsampleH=224, upsampleW=224, isTrain=False, isSingleStream=False): |
| 429 | ''' |