Video interpolation with Super SloMo

This TF2 implementation of Super SloMo is heavily inspired by another PyTorch implementation.

The following packages are required

• tensorflow
• tensorflow-addons
• opencv-python

Training

To train the models, run the following command

python train.py

with the following arguments:

• -v: path to the folder containing the training videos (don't specify if the dataset has already been generated)
• -d: path to the dataset containing the decompressed videos
• -w: path to the weights (this folder contains the checkpoints of the models, saved after each epoch)
• -e: number of epochs
• -b: batch size (default: 7)
• -l: learning rate (default: 0.00002)
• -i: in between frames between the high and low frame-rate videos (default: 7)
• -c: crop size of the random crop used for data augmentation if specified (optional)
• -p: the last checkpoint to load the weights from (optional)
• -s: the number of samples per sequence, used to avoid over-fitting on longer sequences (default: 50)

The weights used are saved into weights. Note: the weights of a pre-trained VGG16 are required for the training phase and should be placed in the weights folder as weights_vgg16.h5. A pre-trained model and the weights of a VGG16 can be found here.

Up-sample

To up-sample a video, run the following command:

python upsample.py

with the following arguments

• -i: the path to the input video
• -o: the path to the output video
• -f: the temporary frame folder (leave unspecified if you don't need to retrieve the frames)
• -mf: the number of frame to extract from the input video
• -p: the model checkpoint to use
• -b: batch size
• -r: the up-sampling rate
• -fps: the frame rate of the output video