Video Frame Interpolation with Densely Queried Bilateral Correlation
Authors: Chang Zhou, Jie Liu, Jie Tang, Gangshan Wu
IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments show that our approach enjoys higher accuracy and less inference time than the state-of-the-art. Source code is available at https://github.com/kinoud/DQBC. |
| Researcher Affiliation | Academia | Chang Zhou , Jie Liu , Jie Tang and Gangshan Wu State Key Laboratory for Novel Software Technology, Nanjing University, China zhouchang@smail.nju.edu.cn, {liujie, tangjie, gswu}@nju.edu.cn |
| Pseudocode | No | The paper includes architectural diagrams (Figure 2, 3, 4, 5) but does not provide any explicitly labeled 'Pseudocode' or 'Algorithm' blocks. |
| Open Source Code | Yes | Source code is available at https://github.com/kinoud/DQBC. |
| Open Datasets | Yes | We use the training set of the Vimeo90K dataset proposed by [Xue et al., 2019] as our training set. The training dataset contains 51,312 triplets with a resolution of 448 256. ... UCF101. The UCF101 dataset was originally proposed by [Soomro et al., 2012] ... Middlebury. Proposed by [Baker et al., 2011] ... SNU-FILM. Proposed with [Choi et al., 2020]. |
| Dataset Splits | No | The paper mentions 'The training set of the Vimeo90K dataset... as our training set' and 'The evaluation set of Vimeo90K dataset... contains 3,782 triplets', but it does not explicitly specify a separate 'validation' dataset split for hyperparameter tuning during training. |
| Hardware Specification | Yes | We use a total batch size of 64 distributed over 4 Tesla V100 GPUs for 510 epochs. The inference time is recorded as the average inference time per sample on the Vimeo90K evaluation set and is tested on the same GTX1080-Ti GPU. |
| Software Dependencies | No | The paper mentions that the network is optimized by Adam W and MACs are calculated using 'thop', but it does not provide specific version numbers for any programming languages, deep learning frameworks (e.g., PyTorch, TensorFlow), or other key software libraries. |
| Experiment Setup | Yes | Our network is optimized by Adam W [Loshchilov and Hutter, 2017] with weight decay 10^-4 on 256x256 patches. The patches are randomly cropped from the training set and are randomly flipped, rotated and reversed. We use a total batch size of 64 distributed over 4 Tesla V100 GPUs for 510 epochs. The training process lasts for about three days. The memory consumption for each GPU is about 18 GB. The learning rate gradually decays from 2x10^-4 to 2x10^-6 using cosine annealing during the training process. |