Results 2 comments of xuduo35

I setup a project based on lucidrains's code and TuneAVideo, train from scratch using 2M webvid videos on 2 RTX3090. https://github.com/xuduo35/MakeLongVideo Some results: ![image](https://github.com/xuduo35/MakeLongVideo/raw/main/samples/a%20video%20of%20Earth-DzP1ma.gif) ![image](https://github.com/xuduo35/MakeLongVideo/raw/main/samples/a%20cat%20eating%20foo-IR47a4.gif) ![image](https://github.com/xuduo35/MakeLongVideo/raw/main/samples/a%20glass%20bead%20fal-Uxxg0y.gif)