本仓库是《基于Transformer和增强可变形可分离卷积的多视频插帧方法》的论文代码
以下为运行代码所需要的依赖:
- python==3.7.6
- pytorch==1.5.1
- cudatoolkit==10.1
- torchvision==0.6.1
- cupy==7.5.0
- pillow==8.2.0
- einops==0.3.0
- opencv-python
- timm
- tqdm
Baidu Cloud: kele
python main.py --model TAE_MVFI_s --dataset vimeo90K_triplet --data_root <dataset_path> --batch_size 8 --num_workers 32
- Vimeo90K triplet
python test.py --model TAE_MVFI_s --dataset vimeo90K_triplet --data_root <dataset_path> --load_from checkpoints/TAE_MVFI_s/model_best.pth
- UCF101
python test.py --model TAE_MVFI_s --dataset ucf101 --data_root <dataset_path> --load_from checkpoints/TAE_MVFI_s/model_best.pth
- DAVIS
python test.py --model TAE_MVFI_s --dataset Davis --data_root <dataset_path> --load_from checkpoints/TAE_MVFI_s/model_best.pth
- 单次插值:
python interpolate_demo.py --model TAE_MVFI_s --load_from checkpoints/TAE_MVFI_s/model_best.pth
- 多次插值:
python interpolate_demo1.py --model TAE_MVFI_s --load_from checkpoints/TAE_MVFI_s/model_best.pth
python main.py --model TAE_MVFI_m --dataset vimeo90K_septuplet --data_root <dataset_path> --batch_size 8 --num_workers 32
- Vimeo90K septuplet
python test.py --model TAE_MVFI_m --dataset vimeo90K_septuplet --data_root <dataset_path> --load_from checkpoints/TAE_MVFI_m/model_best.pth
- UCF101
python test.py --model TAE_MVFI_m --dataset ucf101 --data_root <dataset_path> --load_from checkpoints/TAE_MVFI_m/model_best.pth
- DAVIS
python test.py --model TAE_MVFI_m --dataset Davis --data_root <dataset_path> --load_from checkpoints/TAE_MVFI_m/model_best.pth
- 指定插值中间帧的数量:
python interpolate_demo1.py --model TAE_MVFI_m --load_from checkpoints/TAE_MVFI_m/model_best.pth --inter_num 5
- 指定插值中间帧对应的时间步:
python interpolate_demo2.py --model TAE_MVFI_m --load_from checkpoints/TAE_MVFI_m/model_best.pth --times 0.1,0.3,0.5,0.7,0.9
本论文代码借鉴了以下论文开源代码,在此致谢: