Code for the paper "DeepRemaster: Temporal Source-Reference Attention Networks for Comprehensive Video Enhancement". http://iizuka.cs.tsukuba.ac.jp/projects/remastering/
Hello！I have a question about Inference Time .
I learned in the paper that for a 528 × 400-px input video per frame with a 1080Ti GPU, with 4ms corresponding to the restoration stage. But I tested the inference time under the same conditions is 100ms.
So, I want to understand if you have done some optimizations, thanks.
Hello！I have a question about Inference Time . I learned in the paper that for a 528 × 400-px input video per frame with a 1080Ti GPU, with 4ms corresponding to the restoration stage. But I tested the inference time under the same conditions is 100ms. So, I want to understand if you have done some optimizations, thanks.