$\rm{[MindSpore-phase3]}$ $AMT$

本项目包含了以下论文的mindspore实现：

AMT: All-Pairs Multi-Field Transforms for Efficient Frame Interpolation
Zhen Li^*, Zuo-Liang Zhu^*, Ling-Hao Han, Qibin Hou, Chun-Le Guo, Ming-Ming Cheng
(* denotes equal contribution)
Nankai University
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023

[Paper] [Project Page] [Web demos]

文章官方版本仓库链接: https://github.com/Srameo/LED

目前已经完成部分代码的mindspore转化

正在进行中的工作

• 完整代码的mindspore实现

🔧 Dependencies and Installation

python 3.8
cuda 11.6
MindSpore Nightly https://www.mindspore.cn/install/

1. Clone and enter the repo:

   git clone https://github.com/Men1scus/LED_MindSpore.git 
   cd LED_MindSpore

2. Simply run the install.sh for installation! Or refer to install.md for more details.

   We use the customized rawpy package in ELD, if you don't want to use it or want to know more information, please move to install.md

   bash install.sh
   pip install mindspore-dev -i https://pypi.tuna.tsinghua.edu.cn/simple

3. Activate your env and start testing!

   conda activate LED-ICCV23