Tone

Thanks for your interest and feedback. The provided IFRNet, IFRNet-L and IFRNet-S trained on Vimeo-90K have the same output frame when changing `embt`, since the convolution weight which multiple `embt`...

For frame interpolation on videos, you can refer to #9. Thanks.

Thanks for your question. I think the total training iterations should keep the same whether you use 4 or 8 GPUs. You can try to double the training epochs when...

The LiteFlowNet model used in IFRNet is from [https://github.com/sniklaus/pytorch-liteflownet](https://github.com/sniklaus/pytorch-liteflownet).

For your problems, I have following suggestions: We have tested our pre-trained checkpoints of IFRNet for both 2x and 8x frame interpolation. If your PSNR is only around 15~16 dB,...

In fact, I have developed and deployed this repository on different environment including PyTorch 1.3.0 and PyTorch 1.9.1. Since the IFRNet is concise and does not depend on complex modules,...

If replacing the teacher flow network as RAFT, the frame interpolation accuracy of IFRNet will drop. For the reason, please refer to the Task-Oriented Flow Distillation Loss in our paper....

You can refer to the paper Video Enhancement with Task-Oriented Flow (IJCV 2019). DVF is trained on the same Vimeo90K dataset while LiteFlowNet is not. Therefore using DVF's optical flow...

Thanks for the feedback. This may be a bug in higher version of PyTorch.

Thanks for your interest. I calculate the flow magnitude for each pixel in the image and get the flow magnitude map of shape (batch, 1, h, w), then the entire...