| from utils.flolpips import Flolpips | |
| import torch | |
| device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') | |
| eval_metric = Flolpips().to(device) | |
| batch = 8 | |
| I0 = torch.rand(8, 3, 256, 448).to(device) | |
| I1 = torch.rand(8, 3, 256, 448).to(device) | |
| frame_dis = torch.rand(8, 3, 256, 448).to(device) | |
| frame_ref = torch.rand(8, 3, 256, 448).to(device) | |
| flolpips = eval_metric.forward(I0, I1, frame_dis, frame_ref) | |
| print(flolpips) |