AnimeSR: Learning Real-World Super-Resolution Models for Animation Videos

Authors: Yanze Wu, Xintao Wang, GEN LI, Ying Shan

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental 5 Experiments 5.1 Training Details 5.2 Comparisons with Previous Methods 5.3 Ablation Studies and Discussions
Researcher Affiliation Industry 1ARC Lab, Tencent PCG 2Platform Technologies, Tencent Online Video {yanzewu, xintaowang, enochli, yingsshan}@tencent.com
Pseudocode No The paper describes methods in prose and with architectural diagrams, but does not include structured pseudocode or algorithm blocks.
Open Source Code Yes Codes and models are available at https://github.com/Tencent ARC/Anime SR.
Open Datasets Yes We compare models trained with our AVC dataset and ATD-12K dataset.
Dataset Splits No The paper defines AVC-Train and AVC-Test sets, but does not explicitly mention a separate validation dataset split.
Hardware Specification Yes All the training is performed with Py Torch on four NVIDIA A100 GPUs in an internal cluster.
Software Dependencies No All the training is performed with Py Torch on four NVIDIA A100 GPUs in an internal cluster. No specific version numbers for software dependencies are provided.
Experiment Setup Yes We use the Adam optimizer [25] with a learning rate of 2 10 4 for the first stage and a learning rate of 1 10 4 for the second stage. We set the batch size per GPU, frame sequence length, and patch size of the HR frames to 4, 15, and 256, respectively.