FILM
Frame interpolation model for large-scale action scenes
CommonProductImageImageVideo
Frame Interpolation (Frame Interpolation) is a high-quality frame interpolation neural network model. This model uses a unified single-network method, does not require additional pre-trained networks, such as optical flow or depth networks, but still achieves state-of-the-art results. The model uses a multi-scale feature extractor to share the same convolutional weights at different scales. The model is trained only using frame trios.
FILM Visit Over Time
Monthly Visits
2224288
Bounce Rate
35.64%
Page per Visit
6.7
Visit Duration
00:07:28