文章目录

• 第一波
• • linux下实现DAIN的方法
  • windows使用DAIN-APP
• 第二波来了

先提供两个 址
github项目下载
DAIN-APP(windows)官方下载地址

关于DAIN的介绍，大家可以点击这条链接
让电影动漫统统变丝滑，480帧也毫无卡顿，交大博士生开源插帧软件DAIN

本人研究方向也不在 视频插帧这个方向上，只是想用一用最新的深度学习视频插帧算法。DAIN在linux下配置十分复杂，复现比较困难，贴上github教程：

第一波

linux下实现DAIN的方法

Installation

Download repository:

Before building Pytorch extensions, be sure you have :

Generate our PyTorch extensions:

Generate the Correlation package required by PWCNet:

Testing Pre-trained Models

Make model weights dir and Middlebury dataset dir:

Download pretrained models,

and Middlebury dataset:

preinstallations:

We are good to go by:

The interpolated results are under , where the is used to distinguish different runnings.

Downloading Results

Our DAIN model achieves the state-of-the-art performance on the UCF101, Vimeo90K, and Middlebury (eval and other). Download our interpolated results with:

Slow-motion Generation

Our model is fully capable of generating slow-motion effect with minor modification on the network architecture. Run the following code by specifying to generate x4 slow-motion effect:

or set to or as follows

to generate x8 and x10 slow-motion respectively. Or if you would like to have x100 slow-motion for a little fun.

You may also want to create gif animations by:

Have fun and enjoy yourself!

windows使用DAIN-APP

打开官方 址选一个版本下载下来。我是3070的显卡，所以看到这个适配30×0的更新版本，毫不犹豫就选他了：

没有后续了，等了好久之后 错内存不够了。只有把视频拆成多个视频进行处理了。