PCD的输出设点
Use load_from_local loader
[ ] 0/400, elapsed: 0s, ETA:
400 共400个视频帧
torch.Size([1, 64, 180, 320])
torch.Size([1, 64, 180, 320])
torch.Size([1, 64, 180, 320])
torch.Size([1, 64, 180, 320])
torch.Size([1, 64, 180, 320])
[ ] 1/400, 0.6 task/s, elapsed: 2s, ETA: 644s
torch.Size([1, 64, 180, 320])
torch.Size([1, 64, 180, 320])
torch.Size([1, 64, 180, 320])
torch.Size([1, 64, 180, 320])
torch.Size([1, 64, 180, 320])
[ ] 2/400, 1.1 task/s, elapsed: 2s, ETA: 376s
……
……
……
处理五帧PCD做对齐操作,得到一帧的处理好的图片
TSA的输出设点
Use load_from_local loader
[ ] 0/400, elapsed: 0s, ETA:
400
torch.Size([1, 64, 180, 320])
[ ] 1/400, 0.7 task/s, elapsed: 1s, ETA: 562s
torch.Size([1, 64, 180, 320])
[ ] 2/400, 1.2 task/s, elapsed: 2s, ETA: 335s
torch.Size([1, 64, 180, 320])
[ ] 3/400, 1.5 task/s, elapsed: 2s, ETA: 257s
……
……
……
处理一帧的TSA融合操作,得到一帧处理好的图像
3、
对齐后的特征
# print('aligned_feat:',aligned_feat.size())
# aligned_feat: torch.Size([1, 5, 64, 180, 320])
融合后的特征
# print('feat:',feat.size())
# feat: torch.Size([1, 64, 180, 320])
Use load_from_local loader
[ ] 0/400, elapsed: 0s, ETA:
400
aligned_feat: torch.Size([1, 5, 64, 180, 320])
feat: torch.Size([1, 64, 180, 320])
[ ] 1/400, 0.7 task/s, elapsed: 1s, ETA: 562s
aligned_feat: torch.Size([1, 5, 64, 180, 320])
feat: torch.Size([1, 64, 180, 320])
[ ] 2/400, 1.2 task/s, elapsed: 2s, ETA: 335s
……
……
……